TWI871681B - System and method for presenting three-dimensional content and three-dimensional content calculation apparatus - Google Patents

Abstract

A system and a method for presenting three-dimensional content and a three-dimensional content calculation apparatus are provided. In the method, the calculation apparatus receives a request for presentation content including one or more images from a user device, receives the presentation content from a content delivery network according to the request, processes the images using a first machine-learning model to generate a first predicted result, processes the images using multiple machine-learning models to generate at least a second predicted result and a third predicted result, selects a second machine-learning model from the machine-learning models based on a comparison of the first predicted result with the at least the second predicted result and the third predicted result, processes the images using the second machine-learning model and sends a processing result to the user device. Accordingly, the user device generates a three-dimensional presentation of the presentation content.

Description

Three-dimensional content representation system, method and three-dimensional content calculation device

The present invention relates to a content representation system and method, and in particular to a three-dimensional content representation system, method and three-dimensional content computing device.

The learning ability of a deep neural network can be positively correlated with the number of parameters and the relative accuracy of the parameters. Therefore, for general or non-specific input data sets, a large deep neural network (e.g., a deep neural network with a number of parameters greater than the number of thresholds) can achieve high accuracy. However, large deep neural networks may be accompanied by many disadvantages. For example, large deep neural networks may take up a large amount of memory, consume a large amount of processing resources, take time to process the data set (which can be a problem for real-time operation), use a large training data set to train a large deep neural network to the necessary accuracy, take a long time to train, etc. Therefore, although large deep neural networks can be trained to acceptable accuracy for some datasets, they may not be applicable in many cases.

This article describes a three-dimensional content representation system, which includes a content delivery network, a client device, and a computing device. The content delivery network provides representation content including one or more images. The client device sends a request for the representation content. The computing device is connected to the content delivery device and the client device, and is configured to: receive the request from the client device; receive the representation content from the content delivery network according to the request; use a first machine learning model to process one or more images in the representation content to generate a first prediction result; use multiple machine learning models to process the images to generate at least a second prediction result and a third prediction result, wherein wherein the first machine learning model is larger than the plurality of machine learning models; selecting a second machine learning model from the plurality of machine learning models based on a comparison of the first prediction result with at least the second prediction result and the third prediction result; and processing the image using the second machine learning model and transmitting the processing result to the client device, wherein the client device uses the processing result to generate a three-dimensional representation of the content.

The present invention discloses a three-dimensional content representation method, which is applicable to a three-dimensional content representation system including a content delivery network, a client device and a computing device. The method includes the following steps: the computing device receives a request for a representation content from the client device, the representation content including one or more images; the computing device receives the representation content from the content delivery network according to the request; the computing device processes one or more images in the representation content using a first machine learning model to generate a first prediction result; the computing device processes the images using multiple machine learning models to generate at least a second prediction result and a second prediction result. three prediction results, wherein the first machine learning model is greater than the multiple machine learning models; the computing device selects a second machine learning model from the multiple machine learning models based on a comparison of the first prediction result with the at least second prediction result and the third prediction result; the computing device processes the image using the second machine learning model and transmits the processing result to the client device; and the client device uses the processing result to generate a three-dimensional representation of the content.

The present invention discloses a three-dimensional content computing device, which includes a communication interface, a storage device and a storage device. The communication interface is used to communicate with a client device and a content delivery network. The storage device is used to store instructions for operations performed by a processor. The processor is coupled to the communication interface and the storage device, and is configured to access and execute instructions stored in the storage device to receive a request for a representation content from the client device through the communication interface, and the representation content includes one or more images, and receives the representation content from the content delivery network through the communication interface according to the request, uses a first machine learning model to process one or more images in the representation content to generate a first prediction result, and uses multiple machine learning models to process the images to generate at least a second prediction result and a third prediction result, wherein the first machine learning model is larger than the multiple machine learning models, selecting a second machine learning model from the multiple machine learning models based on a comparison of the first prediction result with the at least second prediction result and the third prediction result, and processing the one or more images using the second machine learning model, and transmitting the processing result to the client device, so that the client device uses the processing result to generate a three-dimensional representation of the representation content.

These exemplary examples are not mentioned to limit or define the present disclosure, but to help understand the present disclosure. Additional embodiments are discussed in the specific embodiments, and further description is provided in the specific embodiments.

The learning ability of a machine learning model can be related to the number of parameters or the number of layers of the model. Increasing the learning ability (e.g., by increasing the number of parameters or the number of layers) can enable the machine learning model to learn from a wider set of data. For example, increasing the number of parameters of a classifier can increase the number of classifications that the classifier can reliably distinguish. Increasing the number of parameters or the number of layers of a machine learning model can also increase the processing cost of executing the model (e.g., processing load, execution time, training time, etc.), which may make the machine learning model inoperable under certain conditions (e.g., such as real-time operation, etc.).

Methods and systems for machine learning model selection for discrete processing tasks are described herein. Multiple small machine learning models can be instantiated and trained to process an input data set in place of a large machine learning model. Because each small machine learning model can include fewer parameters or layers than the large machine learning model, the small machine learning models can be configured to achieve the same degree of accuracy as the large machine learning model for a given portion of the input data set. Each small machine learning model can be configured to process a specific input data set (e.g., a data set including specific characteristics, etc.), or can be configured to produce specific outputs (e.g., a subset of possible outputs that the large machine learning model is configured to produce, etc.). The multiple small machine learning models can be configured together to process the same input data set that the large machine learning model is configured to process with similar accuracy and/or loss as the large machine learning model. However, the small machine learning models having fewer parameters or layers than the large machine learning model can operate more efficiently (e.g., using fewer processing resources to store and/or execute, using smaller training data sets for training, faster training times, etc.).

For example, a large classifier can be configured to classify an input image according to multiple different categories based on objects within the input image. A first small machine learning model can be instantiated to classify the input image according to a subset of the different categories, and a second machine learning model can be instantiated to classify the input image according to the remaining different categories. Alternatively or additionally, the first machine learning model can be instantiated to classify input images characterized by natural lighting (e.g., daylight, etc.), and the second machine learning model can be instantiated to classify input images characterized by synthetic lighting (e.g., flash, incandescent, fluorescent, etc.).

In some examples, a large machine learning model can be compressed into small machine learning models. Compressing the machine learning model can reduce the number of parameters or the number of layers, which can make the compressed machine learning model suitable for processing a portion of the input data set that the large machine learning model can process. Once compressed, the multiple small machine learning models can be instantiated by training the compressed machine learning models using different training data sets. Each small machine learning model can be trained to process a range of input data sets that the corresponding large machine learning model is expected to process.

Once trained, the large machine learning model and multiple small machine learning models can be used to process an input data set. For any input data set, a model selector can determine which machine learning model (among the large machine learning model and the multiple small machine learning models) should process a particular input data set. In some examples, the model selector can sample the input data set to generate a test feature vector, and pass the test feature vector as an input to the machine learning model to generate a corresponding test output. For a deep neural network (DNN), one or more initial layers of the DNN can operate as a feature extractor. The test output from the large machine learning model can be labeled as a pseudo ground truth (e.g., assumed to be true). The model selector can then compare the test output from each small machine learning model to the test output from the large machine learning model. In some examples, the model selector can use an accuracy metric or loss function (e.g., accuracy, precision, area under the curve, logarithmic loss, F1 score, weighted human disagreement rate, cross entropy, mean absolute error, mean square error, etc.). The model selector can identify a specific small machine learning model among the plurality of small machine learning models that has the highest accuracy measure or the lowest loss (according to the loss function). The model selector can then use the specific small machine learning model to process the remainder of the specific input data set.

In an illustrative example, monocular depth estimation may be performed by one or more machine learning models (e.g., DNN or similar models) used for various computer vision operations (e.g., but not limited to, classification, semantic segmentation, object detection, instance segmentation, depth estimation, etc.) (e.g., for autonomous driving of self-driving cars, virtual reality, augmented reality, three-dimensional (3D) analogy, object acquisition, etc.). A computing device may instantiate (define and train) a large DNN to process various images, video frames, or video clips to generate a depth map or an inverse depth map using monocular depth estimation. The depth map may represent each pixel of the image as a distance (e.g., a real number) between the location in the environment represented by the pixel and the camera. Multiple small DNNs may also be trained to process various input images, video frames, and/or video clips. In some examples, each small DNN may be generated by compressing a large DNN and training the compressed large DNN.

The computing device may receive a plurality of images. The plurality of images may be different images, images extracted from video frames, images extracted from video clips, or the like. The images may be received from a content delivery network, a client device, another computing device, a camera (e.g., such as a live camera stream or previously stored images captured from a camera), a server, and the like. Alternatively, the computing device may receive the images by extracting the images from a video clip stored in a memory of the computing device.

The computing device may select one or more images from the plurality of images. In some embodiments, the computing device samples the plurality of images to export the one or more images.

The computing device may process the one or more images using the large DNN to generate a first prediction result corresponding to the output of the large DNN. For example, the computing device may generate a feature vector using the one or more images, and the feature vector may be passed as an input to the large DNN. The large DNN may process the feature vector and output a first prediction result (e.g., a depth map or a reverse depth map, etc.). In some examples, the computing device may treat the first prediction result as a false value.

The computing device may process the one or more images using the multiple small DNNs to generate additional prediction results. For example, the first small DNN may process the one or more images to generate a second prediction result, the second small DNN may process the one or more images to generate a third prediction result, and so on. Each of the small DNNs may be smaller than the large DNN (e.g., fewer parameters and/or layers, etc.).

The computing device may select a small DNN from the plurality of small DNNs based on a comparison of the first prediction with the second prediction, the third prediction, etc. In some examples, the computing device may use one or more accuracy metrics and/or loss functions to compare the second prediction, the third prediction, etc. relative to the first prediction. For example, since the prediction includes a depth map or an inverse depth map (e.g., representing each pixel as a real distance from the camera), a loss function may be used to determine the difference between the first prediction (which is marked as a false value) and the second prediction, the first prediction, the third prediction, etc. Examples of loss functions include, but are not limited to, adaptive robust loss, mean squared error, mean absolute error, cross entropy, weighted human disagreement rate (WHDR), combinations thereof, or similar functions. The computing device may select a particular small DNN with the highest accuracy metric, lowest error, lowest loss, etc.

The computing device may then process the plurality of images using a specific small DNN to generate a depth map or a reverse depth map from the plurality of images. In some examples, the computing device may process each of the plurality of images. In other examples, the computing device may process a portion of the plurality of images by sampling the plurality of images. For example, the computing device processes every nth image of the plurality of images.

In some examples, the model selection process may be repeated to ensure that a particular small DNN remains the most efficient small DNN for processing the plurality of images. The model selection process may be re-executed at regular intervals upon detection of an event, upon detection of user input, after a predetermined number of instances of a particular small DNN are executed therein, upon detection of a change in one or more characteristics of the plurality of images (e.g., such as a change in average pixel value, etc.), in a combination of the above, or in a similar situation. The computing device may continuously ensure that the most efficient small DNN is used to process the plurality of images.

The model selection process can be applied to a variety of machine learning models to determine how to efficiently process completely different sets of data. In this way, the techniques described in this article can be applied to deep neural networks (as described previously) as well as any other type of machine learning model.

1 is a block diagram of an exemplary system for selecting a machine learning model configured to process disparate data sets according to aspects of the present disclosure. The computing device 104 can be configured to process disparate data sets for nearby devices (e.g., such as devices operating within the same network) and/or remote devices (e.g., such as devices operating within other networks, etc.). The computing device 104 may include a central processing unit (CPU) 108, a memory 112 (e.g., volatile memory (e.g., random access memory, etc.) and non-volatile memory (e.g., flash memory, hard disk, etc.)) connected via a bus or similar device, an input/output interface 116, a network interface 120, and a data processor 124. In some implementations, the computing device 104 may include additional components or fewer components.

The input/output interface 116 may include one or more hardware and/or software interfaces configured to receive data from and/or transmit data to one or more devices 132 connected to the computing device 104, such as, but not limited to, a display device, a keyboard and mouse, a sensor, a peripheral device, a media streaming device, an augmented reality device, a virtual reality device, and/or the like. In an illustrative example, a first device of the one or more devices 132 may be a virtual reality display device configured to project a three-dimensional representation of a media (e.g., a video, a video game, one or more images, etc.). If the media does not include three-dimensional data (e.g., the media is two-dimensional, etc.), the computing device 104 may perform monocular depth estimation using the data processor 124 to generate a depth map and generate a three-dimensional representation of the media from the depth map. The computing device 104 may then transmit the three-dimensional representation of the media to the virtual reality display via the input/output interface 116. One or more devices 132 can be connected via a wired connection (e.g., universal serial bus (USB) type A, USB type B, or USB type C; high-definition multimedia interface (HDMI); digital visual interface (DVI); DisplayPort; etc.) or a wireless connection (e.g., such as but not limited to wireless fidelity (Wi-Fi), Bluetooth, Zigbee, Z-wave, infrared, ultra-wideband, etc.).

The network interface 120 may enable connection to one or more remote devices via a network 128 (e.g., the Internet, a local area network, a wide area network, a cloud network, etc.). In some examples, the computing device 104 may receive a request to process data using the data processor 124 via the network interface 120. Once the request is received, the computing device 104 may store the data in the memory 112, process the data using the data processor 124, and transmit the output to the requesting device (or one or more other devices) via the network 128. Alternatively or additionally, the output may be presented via one or more devices 132. In some examples, the data processor 124 may process the received data in real time. In these examples, the data processor 124 may process the streamed data as it is received (via the network interface 120 or the input/output interface 116), or may store a portion of the stream in a buffer in the memory 112 and process a portion of the streamed data stored in the buffer whenever the buffer is full.

In some implementations, the data processor 124 may be a separate component of the computing device 104 that is connected to the CPU 108, the memory 112, the input/output interface 116, and the network interface 120 via a bus. The data processor 124 may be configured to operate within the computing device 104 or may operate independently of the computing device 104. For example, the data processor 124 may be an application specific integrated circuit (ASIC), a field programmable logic gate array, a mask programmable gate array, a microcontroller, or the like that is configured to process instructions stored in a memory of the data processor 124. Alternatively, data processor 124 may be a non-volatile memory (either as a separate component connected to the bus or as a subcomponent of memory 112) configured to process various data sets, the non-volatile memory storing instructions. The instructions may be executed by CPU 108 (and/or other components of computing device 104).

The data processor 124 may include: a model selector 136 configured to select a particular machine learning model for processing a particular data set; a feature extractor 140 configured to generate an input feature vector for a selected machine learning model (e.g., a model that does not require an additional feature extractor); training data 144 for storing training data for the machine learning model; a large machine learning (ML) model 148; and one or more small machine learning models (e.g., such as small ML model 1 152 to small ML model n 156, where n can be any integer greater than 1).

The data processor 124 may use two or more machine learning models to process various types of data sets. The two or more machine learning models may have different sizes, which allow the data processor 124 to dynamically select the most efficient machine learning model for processing a given data set based on the current state of the data processor 124 and/or the computing device 104, or dynamically switch to a different machine learning model. The two or more machine learning models may include a large machine learning model (e.g., a machine learning model with a number of parameters or a number of layers greater than a threshold) and one or more small machine learning models (e.g., a machine learning model with a number of parameters or a number of layers less than a threshold).

The size of a machine learning model (e.g., the number of parameters of a neural network, the number of layers, etc.) can indicate the learning potential of the machine learning model. Large machine learning models can be trained to process general datasets (e.g., datasets that may not correspond to any taxonomy or may not have any particular common characteristics). For example, a large image classifier trained to classify objects within an image can classify randomly sampled input images (e.g., daylight, indoors, night or low light, objects to be classified are obscured or far away from the camera, objects to be classified are clear and close to the camera, etc.). Small machine learning models may have lower accuracy and/or higher loss when classifying certain types of images. For example, a small image classifier trained to classify objects within an image may classify images that share certain characteristics (e.g., images taken during the day or with a lot of light) and may have lower accuracy or higher loss when classifying images with different characteristics (e.g., images captured at night or in low-light conditions, etc.).

Large machine learning models may have a larger memory footprint and may use more processing resources (e.g., CPU 108, cache or volatile memory, non-volatile memory, bandwidth, etc.) than corresponding small machine learning models. Large machine learning models may also have different training intervals than small machine learning models and execute over longer time intervals, making the use of large machine learning models more complex for time-sensitive operations.

Machine learning models 148-156 may be any type of machine learning model, including but not limited to neural networks, deep neural networks, transformers, classifiers, support vector machines, decision trees, etc. In some examples, machine learning models 152-156 may be generated by compressing a large machine learning model 148 (before, during, or after the large machine learning model 148 is trained). In these examples, the large machine learning model 148 can be compressed by pruning (e.g., removing unnecessary parameters or layers, etc.), quantization (e.g., reducing the memory area occupied by parameters, etc.), knowledge distillation (e.g., training a small machine learning model to simulate a large machine learning model), low-rank factorization, etc.

The large machine learning model 148 may be trained and/or compressed by the data processor 124. One or more small machine learning models 152 to 156 may be generated by compressing the large machine learning model 148 and/or trained by the data processor 124. The model selector 136 may determine a machine learning model of the type of processing task to be performed for the selected processing task. The model selector 136 may pass a training request to the feature extractor 140. The feature extractor 140 may generate a training data set to train a machine learning model of the type of processing task to be performed. The data processor 124 may train a machine learning model to perform one or more operations using training data stored in the training data 144, generated (e.g., programmatically generated by the feature extractor 140 or received from user input), or received from one or more remote devices (e.g., one or more devices 132, one or more remote devices connected via the network 128, etc.). The training data 144 may store data configured to train the machine learning model to process a specific type of input data. For example, training data 144 may store image data so that a machine learning model may be trained to process images (e.g., generate depth maps, classify images, detect objects, etc.). Training data 144 may also store historical data (e.g., data associated with historical executions of machine learning models 148 to 156, etc.), generated data, received data, etc. If one or more small machine learning models 152-156 are to be trained independently of the large machine learning model 148, the feature extractor 140 may generate a training set for the one or more small machine learning models 152-156 based on the type of machine learning model and the size of the machine learning model to be trained. The training data set used to train the one or more small machine learning models 152-156 may be similar to or the same as the training data set used to train the large machine learning model 148.

The feature extractor 140 can use the training data set to train the machine learning models 148 to 156. The machine learning models 148 to 156 can be trained for a predetermined number of iterations at a predetermined time interval, until a target accuracy metric is achieved, until a target loss value (according to one or more loss functions) is achieved, and so on.

The data processor 124 may receive a data set for processing using one or more of the trained machine learning models 148 to 156. The data set may be received via the input/output interface 116, the network interface 120, or stored in the memory 112. The data set may be a discrete data set (e.g., of a definite size and/or length, etc.) or may be a continuous stream (e.g., such as broadcast media, video games, or other media of uncertain size or length, etc.). The model selector 136 may determine which of the machine learning models 148-156 will most efficiently process a received data set (or a portion of a received data set) by sampling the data set, processing the samples using a machine learning model, and determining which machine learning model should process the data set by comparing the results.

The model selector 136 may sample the dataset by extracting a portion of the dataset. The model selector 136 may sample an initial portion of the dataset (e.g., a first number of bits, a first number of images or video frames, a first predetermined second of audio, etc.). Alternatively or additionally, the model selector 136 may obtain a random sample of the dataset by randomly selecting a portion of the dataset, for example, using a random number generator. The model selector 136 may send an identification of the sample and an indication of which machine learning models are to be utilized to the feature extractor 140.

The feature extractor 140 may generate a feature vector for a selected machine learning model (e.g., the large machine learning model 148 and one or more machine learning models 152 to 156 that do not include an internal feature extraction function, etc.). The large machine learning model 148 may be executed using the feature vector from the feature extractor 140 and generate a first output (e.g., a first prediction result). Small machine learning models 1 152 to small machine learning models n 156 may also be executed using the same feature vector (or a feature vector customized by the feature extractor 140 for the corresponding small machine learning model) to generate a second output (e.g., a prediction result from small machine learning model 1 152) to the nth output (from small machine learning model n 156, etc.).

The model selector 136 may compare the first output with the second to nth outputs to determine which of the small machine learning models should be processing the data set. For example, the model selector 136 may mark the first output as true, and then use the first output to measure the accuracy and/or loss of the second to nth outputs to determine the accuracy (e.g., using an accuracy metric, etc.) and/or loss (e.g., using a loss function, etc.) of each small machine learning model 152 to 156 relative to the large machine learning model 148 for the particular data set. The model selector 136 may select the small machine learning model 152 to 156 with the highest accuracy and/or lowest loss. Alternatively, the model selector 136 may measure the second to nth outputs relative to the first output and/or relative to the second to nth outputs to generate a distribution of relative outputs (e.g., the second output relative to the third output, the second output relative to the fourth output, the second output relative to the nth output, etc.), based on which a particular output may be selected as a preferred output over other outputs. Then, a small machine learning model corresponding to the particular output may be selected to process the data set. Alternatively, the model selector 136 may measure each of the second to nth outputs independently of the other outputs to determine the small machine learning model that should process the data set. In this example, the large machine learning model 148 may not be used (eg, the first output may not be generated, etc.). The model selector 136 may measure the output relative to other outputs using any accuracy metric and/or loss function.

Alternatively, model selector 136 may determine that large machine learning model 148 should be used to process the data set instead of small machine learning models 152-156. Model selector 136 may determine that small machine learning models 152-156 may have an accuracy metric that is less than a first threshold and/or a loss function that is greater than a second threshold. Therefore, model selector 136 may select that large machine learning model 148 will be the most efficient machine learning model to use to process the data set. The model selector 136 may select a machine learning model by balancing processing efficiency (e.g., a small machine learning model 152-156 may be more efficient using fewer processing resources) and accuracy (a large machine learning model 148 may sometimes be more accurate than a small machine learning model 152-156).

Model selector 136 may select a small machine learning model as long as the accuracy of the selected small machine learning model relative to the large machine learning model 148 is: 1) greater than other small machine learning models under consideration; and 2) greater than a first threshold. Examples of accuracy metrics and/or loss functions include, but are not limited to, accuracy, precision, area under the curve, log loss, F1 score, weighted human disagreement rate, cross entropy, mean absolute error, mean squared error, or similar functions. Model selector 136 may use the small machine learning model with the highest accuracy and/or lowest loss to begin processing the remainder of the data set.

In some examples, the model selector may perform the machine learning model selection process again during processing of the data set to ensure that the selected small machine learning model is still the most efficient machine learning model for processing the data set. For example, the data processor 124 may process a video stream to produce an estimated depth map for each video frame (or every nth frame, etc.). The first few video frames of the video stream may correspondingly include high-light conditions for which the small machine learning model 1 152 is shown to be the most efficient (based on the execution of the aforementioned model selection process). Subsequent portions of the video stream may include video frames that include low-light conditions for which small machine learning model 1 152 may not be most efficient (e.g., small machine learning model 1 152 may have lower accuracy and/or higher loss when processing low-light video frames). Model selector 136 may use one or more recent inputs to small machine learning model 1 152 to re-execute the model selection process and select one of large machine learning model 148 and/or small machine learning model 2 (not shown) through small machine learning model n 156 to take over processing of the video stream.

The model selection process may be re-executed at regular intervals (e.g., every n video frames, every n seconds, etc.) upon detecting an event, upon receiving user input, upon detecting a change in one or more characteristics of the portion of the data set that is input to the selected small machine learning model (during a particular iteration) and/or an output from the selected small machine learning model (e.g., a change in average pixel value, etc., as in the previous example), an accuracy metric and/or a loss function, a combination of the above, or the like. The model selector 136 may continuously monitor the execution of the selected small machine learning model for a given data set to ensure that the most efficient small machine learning model is being executed.

FIG. 2 illustrates an exemplary three-dimensional content representation system according to various aspects of the present disclosure. In some examples, computing device 104 may operate as a load balancer by providing processing services to one or more client devices (e.g., client device 204). For example, client device 204 may be any processing device, such as but not limited to a desktop or laptop computer, a mobile device (e.g., a smart phone, a tablet computer, etc.), a video game console, a server, etc. Client device 204 may operate processing intensive applications. Client device 204 may use the resources of computing device 104 by transmitting and/or streaming data sets to computing device 104. The computing device 104 may use the data processor 124 to select a small machine learning model that is configured to process the data set to produce an output (or output stream). The computing device 104 may transmit (or stream) the output back to the client device 204.

In other examples, computing device 104 may process data sets that may not be locally processed by client device 204. For example, computing device 104 may operate a virtual reality application configured to present a three-dimensional representation of various media (e.g., movies, video games, analogies, etc.). Computing device 104 may receive content associated with the virtual reality application from content delivery network 212 via network 208. In another example, computing device 104 may receive images from a live camera feed (or images gathered from a live camera feed), and if the content is not already a three-dimensional representation, computing device 104 uses monocular depth estimation to convert the content into a three-dimensional representation. Monocular depth estimation is the process of determining the approximate distance between a surface represented in an image (or video frame) and the camera that captured the image. In some examples, monocular depth estimation may be performed for each pixel in an image (or video frame), resulting in a depth map. The distances may be used to generate a three-dimensional representation of the two-dimensional image. The three-dimensional representation may be used in computer vision applications such as augmented reality, virtual reality, 3D television, video games, mapping three-dimensional environments, simulations, vehicle automation (e.g., self-driving cars), and the like.

Computing device 104 may receive a request for three-dimensional content from client device 204. Computing device 104 may request content from content delivery network 212 and process the content (in real time). In some examples, computing device 104 may transmit content and a depth map generated by a data processor (e.g., such as data processor 124) of computing device 104 to client device 204. Client device 204 may use the content and the depth map to generate a three-dimensional representation of the content for a virtual reality application. Alternatively, client device 204 may receive the content directly from content delivery network 212 and the depth map from computing device 104. Each depth map may be associated with metadata indicating the location of the content corresponding to the depth map. In other examples, computing device 104 may generate 3D representations of content and transmit or stream the 3D representations of content to client device 204. For example, client device 204 may connect to computing device 104 and stream various 3D representations of content generated by computing device 104 from a content delivery network.

FIG3 illustrates a block diagram of an exemplary distributed data processing network according to aspects of the present disclosure. Computing device 104 may operate in a distributed network configured to provide processing services to one or more devices (e.g., client device 204, other devices, servers, networks, etc.). Computing device 104 may include data processor 124 configured to process various data sets. Data processor 124 may use a model selection process for a large machine learning model and one or more small machine learning models to determine the most efficient machine learning model to use when processing a particular data set. The small model selection process can balance or reduce the processing load of the computing device 104 and achieve overall accuracy when selecting a machine learning model to process a particular data set. In some examples, the computing device 104 can operate on multiple large machine learning models and one or more corresponding small machine learning models to achieve parallel processing of similar and/or completely different data sets.

In some examples, computing device 104 may operate as a node in a distributed data processing network. Any number of additional computing devices (e.g., computing device 104-1, computing device 104-2, computing device 104-3, computing device 104-n, etc.) may also operate in the distributed data processing network. Computing devices 104 and each of computing devices 104-1 to 104-n may include a data processor (e.g., such as data processor 124) having a large machine learning model and one or more small machine learning models, and a model selector configured to identify the most efficient small machine learning model that can process a given data set with a threshold accuracy and/or loss.

Computing device 104 may also include a load balancer configured to identify a specific computing device that can process a specific data set. For example, client device 204 may transmit a request with an identification of a specific data set to be processed to computing device 104. The load balancer may select a computing device that can process the specific data set from computing device 104 and computing devices 104-1 to 104-n. The load balancer may select a computing device using one or more characteristics of a particular data set and computing devices 104 and computing devices 104-1 to 104-n, the one or more characteristics being, for example, but not limited to, the processing load of each corresponding computing device, the data type of the particular data set, the expected output, the type of data that can be processed by the corresponding computing device, the network bandwidth, the transmission path (e.g., for transmitting the particular data set to each corresponding computing device and for transmitting the output back to the client device 104, etc.), the accuracy and/or loss of a machine learning model configured to process the particular data set (as determined using the model selection process as previously described), combinations thereof, or the like. In some examples, the one or more features may be weighted, where the weights are continuously adjusted based on the state of the distributed data processing network. For example, features corresponding to the capabilities of the computing device based on a particular data set may be given high weights to ensure that the selected computing device is capable of processing the particular data set (e.g., if the particular data set includes image data, the small machine learning model of the selected computing device is trained to process image data, etc.). Other features may be weighted to balance the processing load on the distributed data processing network.

In some instances, a data set may be processed by more than one computing device. For example, a large data set or media stream may be processed in discrete segments (e.g., per image or video frame, per n-second chunk, per n-bits of data, etc.). Computing device 104 may generate a sequence of feature vectors. Alternatively, for real-time operation, the computing device may generate feature vectors as the computing device receives the data. Each feature vector may be associated with a sequence identifier that indicates the portion of the data set from which the feature vector originates. Computing device 104 may then transmit the feature vector to a computing device selected by the load balancer to process the feature vector and generate an output. The output from the selected computing device and the identifier of the feature vector and/or the sequence identifier may be received by the computing device 104. The computing device 104 may then assemble the output sequence (when processing non-real-time data) from the output received from the computing devices processing the data set, or transmit each output to the client device 204 as it is generated. By distributing the data set across the computing devices of the distributed data processing network, the computing device 104 may reduce the processing load of the distributed data processing network, reduce processing latency by processing portions of the data set in parallel, maintain the accuracy of the data set being processed, and so on.

In other examples, computing device 104 may transmit the selected small machine learning model to client device 204 to enable client device 204 to process the data set locally. In these examples, a data processor of computing device 104 may use a sample of the data set to perform a model selection process to identify a specific small machine learning model that can process the data set with a threshold accuracy or loss (as previously explained). Computing device 104 may then transmit the selected small machine learning model to client device 204. The client device may process the remainder of the data set using the selected small machine learning model.

FIG4 illustrates a block diagram of an exemplary model selection process for selecting a small machine learning model for processing a data set in accordance with aspects of the present disclosure. The model selection process may identify a particular machine learning model to be used to process a given data set. A large machine learning model 408 (e.g., having a number of parameters and/or a number of layers greater than a threshold) may be trained to process a general data set. The large machine learning model 408 may be compressed to produce one or more small machine learning models (e.g., small model 1 412, small model 2 416, small model 3 420, small model n 424, etc.). The large machine learning model 408 may be compressed before, during, or after training. The large machine learning model 148 may be compressed by pruning (e.g., removing unnecessary parameters or layers, etc.), quantization (e.g., reducing the memory footprint of parameters, etc.), knowledge distillation (e.g., training small machine learning models to resemble the large machine learning model), low-rank factorization, or by any other compression algorithm. Alternatively, the one or more small machine learning models may be defined and trained independently. Any number of small machine learning models may be generated (either by compression or by independent training), where n is any integer greater than 1.

When a request is received to process a particular data set, the model selection process can begin. The particular data set can be sampled to produce one or more discrete portions of the data set that can be processed by the machine learning models 408 to 424. One or more feature vectors can be exported from each of the one or more discrete portions of the data set. In some examples, a single feature vector can be exported for the machine learning models 408 to 424. In other examples, a feature vector can be exported for each machine learning model, and the feature vector can be customized for the machine learning model (e.g., based on the number of parameters and/or the number of layers of the model, etc.). Each machine learning model 408-424 may be executed using the one or more feature vectors to produce a corresponding model output. Large machine learning model 408 may process the feature vectors to produce model output 428. Small machine learning model 1 412 may process the feature vectors to produce model 1 output 432, and small machine learning model n 424 may process the feature vectors to produce model n output 444, and so on.

At benchmark selection 452, one or more benchmarks may be selected to evaluate model outputs 428-444. Benchmark selection 452 may designate a model output from large machine learning model 408 (e.g., model output 428) as a true value and compare model outputs from small machine learning models 412-424 (e.g., model outputs 432-444) relative to model output 428.

In some examples, the benchmark selection may determine a benchmark based on the data type of the model outputs 428-444. For example, a benchmark for a classifier may be an accuracy measure or error measure that evaluates the output based on a Boolean value (e.g., true/false or correct/incorrect, etc.). A benchmark for a machine learning model that outputs a numerical output (e.g., such as a depth estimation machine learning model that may output a depth map or an inverse depth map, etc.) may be a loss function (e.g., that determines the difference between a control value and the output). The benchmark selection 452 may use one or more benchmarks when evaluating the outputs 428-444.

In some examples, for outputs including depth maps, the benchmark selection 452 may use a weighted human disagreement rate, a mean absolute relative error, a robust function loss, or a similar function. The weighted human disagreement rate uses equal weights (e.g., set to 1) and identifies for each pixel of the output depth map whether the pixel is closer or farther away from the corresponding pixel of the ground truth model output 428. Each pixel of the model output may be replaced with 0 (indicating that the pixel of the model output is closer than the corresponding pixel of the ground truth depth map) or 1 (indicating that the pixel of the model output is farther away than the corresponding pixel of the ground truth depth map). The distribution of 0 and 1 can be used to evaluate the degree to which the model output depth map deviates from the truth. The mean absolute relative error can be used The error is evaluated, where corresponds to the value of pixel I on the depth map being evaluated, corresponds to the ground truth value of pixel i in the ground truth depth map 428, and M corresponds to the total number of pixels in the depth map.

The robust loss function can be represented by d, d represents the predicted disparity (between the true value of a given pixel and the model output), d* represents the true value of the pixel from the ground-truth model output, and M represents the number of pixels in the depth map.

In other examples, any accuracy metric, loss function, error rate, or similar function can be used to evaluate the model output relative to the true model output. Examples of accuracy metrics and/or loss functions include, but are not limited to, accuracy, precision, area under the curve, log loss, F1 score, weighted human disagreement rate, cross entropy, mean absolute relative error, mean square error, or similar functions.

Baseline selection 452 may then identify the model outputs 432 to 444 having the highest accuracy or lowest loss. Model selection 456 may then select the small machine learning model corresponding to the identified model output from the small machine learning models 412 to 424. The remainder of the particular data set may be processed using the selected small machine learning model.

FIG5 is a flow chart of an exemplary process for model selection for monocular depth estimation according to aspects of the present disclosure. Monocular depth estimation may be implemented by one or more machine learning models (e.g., deep neural networks or similar models) that are used for various computer vision operations (e.g., but not limited to classification, semantic segmentation, object detection, instance segmentation, depth estimation, etc.) (e.g., such as for autonomous driving of self-driving cars, virtual reality, augmented reality, 3D analogy, target acquisition, etc.). A model selection process may be performed to select an efficient machine learning model based on the processing resources consumed by the selected machine learning model and the accuracy of the machine learning model when processing a particular data set.

For example, at block 504, a computing device may receive a plurality of images. The plurality of images may be independent images (independent of other images in the plurality of images, etc.), images extracted from a video frame, images extracted from a video clip, or the like. The images may be received from a camera, a content delivery network, a client device, another computing device, a server, etc. Alternatively, the computing device may receive the images by extracting the images from a video clip stored in a memory of the computing device or from a live camera stream. For a live camera stream, the images may be received continuously as they are captured by the camera. Since additional images of the live camera stream will be received over time, once the first image from the live camera stream is received, the computing device is considered to have received the multiple images.

At block 508, the computing device may select one or more images from the plurality of images. For example, the computing device may sample the plurality of images to export the one or more images. For a live camera stream, the one or more images may correspond to the one or more images received from the live camera stream. In some examples, the computing device may randomly sample the plurality of images. In other examples, the computing device may select the one or more images from the plurality of images based on one or more parameters. The one or more parameters may be based on the number and/or order of images in the plurality of images, characteristics of the images (e.g., pixel values (e.g., average red value, green value, blue value and/or pixel brightness value, etc.)), metadata associated with the plurality of images, combinations thereof, or the like. For example, the computing device may sample the plurality of images by selecting images uniformly over the distribution of the plurality of images based on the number of images to be included in the sampling (e.g., such as the first image and the last image when two images are sampled from the plurality of images, etc.).

At block 512, the computing device may process the one or more images using a first machine learning model to generate a first prediction result. For example, the computing device may use the one or more images to generate a feature vector. The feature vector may be passed as an input to the first machine learning model. The first machine learning model may process the feature vector and output a first prediction result. The first machine learning model may be a large machine learning model. The large machine learning model may be a machine learning model having a number of parameters and/or a number of layers greater than a threshold. In some examples, the computing device may mark the first prediction result as a false value that can be used to compare the output of other machine learning models with the first prediction result.

At block 516, the computing device may process the one or more images using multiple machine learning models to generate a second prediction result (e.g., using a first machine learning model from the multiple machine learning models), a third prediction result (e.g., using a second machine learning model from the multiple machine learning models), etc. The multiple machine learning models may be small machine learning models. A small machine learning model may include a number of parameters and/or a number of layers that is less than a threshold. The first machine learning model may include more parameters and/or layers than the multiple machine learning models. The multiple machine learning models may be generated by compressing the first machine learning model (e.g., using pruning, quantization, knowledge extraction, low-rank factorization, etc.) before, during, or after training the first machine learning model. In some examples, the computing device determines whether to train the multiple machine learning models. The computing device may determine to train one or more of the multiple machine learning models using the same training data as used to train the first machine learning model, training data similar to that used to train the first machine learning model, or training data different from that used to train the first machine learning model. Alternatively, the multiple machine learning models can be defined and trained independently (e.g., separately from the first machine learning model). In these examples, the multiple machine learning models can be of the same, similar, or different types as the first machine learning model (e.g., different models, different parameters, different types of layers, different algorithms, different training processes or iterations, etc.).

The computing device may use the same or similar feature vector (exported from the one or more images) that is passed as input to the first machine learning model as input to input the multiple machine learning models to generate a second prediction result, a third prediction result, etc. In other examples, the computing device may customize the feature vectors for the multiple machine learning models. Because the multiple machine learning models have fewer parameters and/or layers, the machine learning models may accept fewer features in the input feature vectors. The computing device may compress the feature vectors (e.g., using any of the aforementioned compression techniques) to reduce the number of input features.

At block 520, the computing device may select a second machine learning model from the plurality of machine learning models based on a comparison of the first prediction to the second prediction and the third prediction, etc. In some examples, the computing device may use one or more accuracy metrics and/or loss functions to compare the second prediction, the third prediction, etc. relative to the first prediction. For example, a depth map may represent each pixel of an image as a distance (e.g., a real number) between a location in the environment represented by the pixel and the camera. The computing device may perform a pixel-by-pixel comparison of each distance value for each pixel of the second prediction relative to each corresponding pixel in the first prediction (e.g., considered to be a true value for comparison purposes). A loss function may be used to determine the comparison between the second prediction result and the first prediction result, the third prediction result and the first prediction result, etc. Examples of loss functions include but are not limited to adaptive robust loss, mean square error, mean absolute error, cross entropy, weighted human disagreement rate (WHDR), combinations thereof, or similar functions. The computing device may select a machine learning model with the highest accuracy metric, the lowest error rate, the lowest loss, etc. from the multiple machine learning models to process the multiple images.

At block 524, the computing device may process the plurality of images using a second machine learning model (e.g., the machine learning model with the highest accuracy or lowest loss selected from block 520, etc.). In some examples, the computing device may process each of the plurality of images. In other examples, the computing device may process a portion of the plurality of images by sampling the plurality of images (e.g., such as every nth image, etc.). Returning to the monocular depth estimation example, the computing device may use the second machine learning model and the plurality of images to generate a sequence of depth maps. Because the second machine learning model is a small machine learning model, monocular depth estimation can be performed in near real time (e.g., using a live camera stream, dynamic or procedurally generated images from a video game, etc.).

In some examples, the model selection process may be repeated to ensure that the second machine learning model remains the most efficient machine learning model for processing the multiple images. The model selection process may be re-executed at regular intervals upon detection of an event, upon detection of user input, after predetermined repetitive operations of the second machine learning model, upon detection of a change in one or more characteristics of the multiple images (e.g., such as a change in average pixel value, etc.), in a combination of the above situations, or in similar situations. The computing device may continuously ensure that the most efficient machine learning model is used to process the multiple images (e.g., a machine learning model with the highest accuracy, lowest error rate, lowest loss, etc. when processing the multiple images).

The model selection process can be applied to a variety of machine learning models to determine how to efficiently process completely different datasets. In this way, the techniques described in this article can be applied to deep neural networks (as described previously) as well as any other type of machine learning model or dataset.

FIG6 illustrates an exemplary computing device according to various aspects of the present disclosure. For example, computing device 600 can implement any of the systems or methods described herein. In some examples, computing device 600 can be an element of a media device or included in a media device. The elements of computing device 600 are shown as being in electrical communication with each other using connectors 606, such as buses. The exemplary computing device 600 includes a processor (e.g., a CPU, a processor, or the like) 604 and a connection 606 (e.g., such as a bus or the like) configured to couple elements of the computing device 600 (e.g., but not limited to, a memory 620, a read only memory (ROM) 618, a random access memory (RAM) 616, and/or a storage device 608) to the processor 604.

The computing device 600 may include a cache 602 of high-speed memory that is directly connected to, adjacent to, or integrated into the processor 604. The computing device 600 may copy data from the memory 620 and/or storage device 608 to the cache 602 for faster access by the processor 604. In this way, the cache 602 may provide a performance boost that avoids delays when the processor 604 is waiting for data. Alternatively, the processor 604 may access data directly from the memory 620, ROM 618, RAM 616, and/or storage device 608. The memory 620 may include various types of homogeneous memory or heterogeneous memory (eg, such as but not limited to, magnetic memory, optical memory, solid-state memory, etc.).

The storage device 608 may include one or more non-transitory computer-readable media, such as volatile memory and/or non-volatile memory. The non-transitory computer-readable media may store instructions and/or data that may be accessed by the computing device 600. The non-transitory computer-readable media may include, but are not limited to, magnetic cassettes, hard-disk drives (HDDs), flash memory, solid-state memory devices, digital versatile disks, cartridges, optical disks, random access memory (RAM) 616, read-only memory (ROM) 618, combinations thereof, or the like.

The storage device 608 may store one or more services (e.g., service 1 610, service 2 612, and service 3 614) executable by the processor 604 and/or other electronic hardware. The one or more services include instructions executable by the processor 604 to perform the following operations: perform operations, such as any of the techniques, steps, processes, blocks, and/or operations described herein; control the operation of devices in communication with the computing device 600; the operation of the processor 604 and/or any special purpose processor; combinations thereof; or similar devices. The processor 604 may be a system on a chip (SOC) that includes one or more cores or processors, buses, memory, clocks, memory controllers, caches, other processor components, and/or the like. Multi-core processors may be symmetric or asymmetric.

The computing device 600 may include one or more input devices 622, which may represent any number of input mechanisms, such as a microphone, a touch screen for graphic input, a keyboard, a mouse, motion input, voice, a media device, a sensor, a combination thereof, or the like. The computing device 600 may include one or more output devices 624 for outputting data to a user. Such output devices 624 may include, but are not limited to, a media device, a projector, a television, a speaker, a combination thereof, or the like. In some examples, a multimodal computing device may enable a user to provide multiple types of input to communicate with the computing device 600. The communication interface 626 may be configured to manage user input and computing device output. The communication interface 626 may also be configured to manage communications with remote devices (e.g., establish connections, receive/send communications, etc.) via one or more communication protocols and/or via one or more communication media (e.g., wired, wireless, etc.).

The computing device 600 is not limited to the components shown in Figure 6. The computing device 600 may include other components not shown and/or may omit the components shown.

The following examples illustrate various aspects of the present disclosure. As used below, any reference to a series of examples will be understood as a separate reference to each of these examples (e.g., "Example 1 to Example 4" should be understood as "Example 1, Example 2, Example 3, or Example 4").

Example 1 is a computer-implemented method, which includes: receiving multiple images; selecting one or more images from the multiple images; processing the one or more images using a first machine learning model to generate a first prediction result; processing the one or more images using multiple machine learning models to generate at least a second prediction result and a third prediction result, wherein the first machine learning model is larger than the multiple machine learning models; selecting a second machine learning model from the multiple machine learning models based on a comparison of the first prediction result with the at least second prediction result and the third prediction result; and processing the multiple images using the second machine learning model.

Example 2 is the computer-implemented method described in any of Examples 1 and Examples 3 to 8, wherein selecting a second machine learning model from the multiple machine learning models includes: generating a first accuracy value for the first prediction result, a second accuracy value for the second prediction result, and a third accuracy value for the third prediction result; comparing the first accuracy value, the second accuracy value, and the third accuracy value, wherein the second machine learning model is selected based on the second accuracy value being higher than the first accuracy value and the third accuracy value.

Example 3 is a computer-implemented method as described in any of Examples 1 to 2 and Examples 4 to 8, wherein selecting a second machine learning model includes: using a loss function to generate a first value of a first prediction result, a second value of a second prediction result, and a third value of a third prediction result; comparing the first value, the second value, and the third value, wherein the second machine learning model is selected based on the second value being higher than the first value and the third value.

Example 4 is the computer-implemented method of any of Examples 1 to 3 and 5 to 8, wherein the second machine learning model is configured to generate a depth estimation map for an image in the plurality of images.

Example 5 is the computer-implemented method of any of Examples 1 to 4 and Examples 6 to 8, wherein the second machine learning model is configured to perform semantic segmentation on an image in the plurality of images.

Example 6 is the computer-implemented method of any of Examples 1 to 5 and Examples 7 to 8, wherein the second machine learning model is configured to perform instance segmentation using an image from the plurality of images.

Example 7 is a computer-implemented method as described in any of Examples 1 to 6 and 8, wherein the multiple machine learning models are deep neural networks.

Example 8 is the computer-implemented method of any of Examples 1 to 7, wherein the first machine learning model includes more layers than each of the plurality of machine learning models.

Example 9 is a system comprising: one or more processors; a non-transitory computer-readable medium storing instructions that, when executed by the one or more processors, cause the one or more processors to perform the operations described in any of Examples 1 to 8.

Example 10 is a non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform the operations described in any of Examples 1 to 8.

The term "computer-readable medium" includes, but is not limited to, portable or non-portable storage devices, optical storage devices, and various other media capable of storing, containing, or carrying instructions and/or data. Computer-readable media may include non-transitory media in which data may be stored in a form that does not include carrier waves and/or electronic signals. Examples of non-transitory media may include, but are not limited to, magnetic disks or tapes, optical storage media (such as compact disks (CDs) or digital versatile disks (DVDs)), flash memory, memory, or memory devices. Computer-readable media may store thereon code and/or machine-executable instructions, which may represent procedures, functions, subroutines, programs, routines, subroutines, modules, software packages, classes, or any combination of instructions, data structures, or programming statements. A code segment may be coupled to another code segment or hardware circuit by passing and/or receiving information, data, arguments, parameters, or memory contents. Information, arguments, parameters, data, etc. may be passed, forwarded, or transmitted by any suitable means, including memory sharing, message passing, token passing, network transmission, or the like.

Portions of this specification describe examples in terms of algorithms and symbolic representations of operations on information. These operations (although described functionally, computationally, or logically) may be implemented by computer programs or equivalent circuits, microcodes, etc. In addition, arrangements of operations may be referred to as modules without loss of generality. The described operations and their associated modules may be implemented in software, firmware, hardware, or any combination thereof.

Any of the steps, operations or processes described herein may be implemented or performed using one or more hardware or software modules alone or in combination with other devices. In some examples, a computer-readable medium storing computer program code may be used to implement a software module, which may be executed by a processor to perform any or all of the steps, operations or processes described.

Some examples may involve an apparatus or system for performing any or all of the steps, operations, or processes described. The apparatus or system may be specially constructed for the desired purpose, and/or the apparatus or system may include a general-purpose computing device that is selectively activated or reconfigured by a computer program stored in a memory of the computing device. The memory may be or include a non-transitory tangible computer-readable storage medium, or any type of medium suitable for storing electronic instructions (which may be coupled to a bus). In addition, any computing system mentioned in the specification may include a single processor or multiple processors.

Although the subject matter has been described in detail with respect to specific examples, it should be understood that variations, modifications, and equivalents of such embodiments can be easily generated by a person skilled in the art based on an understanding of the foregoing. Many specific details are set forth herein to provide a thorough understanding of the claimed subject matter. However, a person skilled in the art will understand that the claimed subject matter can be practiced without these specific details. In other examples, methods, apparatuses, or systems known to a person of ordinary skill in the art have not been described in detail to avoid obscuring the claimed subject matter. Therefore, the present disclosure has been presented for purposes of example and not limitation, and does not exclude the inclusion of such modifications, modifications, and/or additions to the subject matter as would be apparent to a person of ordinary skill in the art.

For clarity of explanation, in some examples, the disclosure may be presented as including separate functional blocks, including functional blocks including devices, device elements, steps or routines in methods implemented in software, or combinations of hardware and software. Additional functional blocks other than those shown in the figures and/or described herein may also be used. For example, circuits, systems, networks, processes, and other elements may be shown as elements in block diagram form to avoid obscuring the embodiments with unnecessary detail. In other examples, to avoid obscuring the embodiments, the well-known circuits, processes, algorithms, structures, and techniques shown may be free of unnecessary detail.

Individual examples may be described herein as processes or methods, which may be depicted as flow charts, flow diagrams, data flow diagrams, structure diagrams, or block diagrams. Although a flow chart may describe an operation as a sequential process, many of the operations may be performed in parallel or simultaneously. In addition, the order of the operations may be rearranged. When the operations of a process are completed, the process is terminated, but may have additional steps that are not shown. A process may correspond to a method, function, procedure, subroutine, subprogram, etc. When a process corresponds to a function, its termination may correspond to returning the function to the calling function or the main function.

The processes and methods according to the above examples may be implemented using computer executable instructions stored in or otherwise obtained from a computer readable medium. Such instructions may include, for example, instructions and data that cause or otherwise configure a general purpose computer, a special purpose computer, or a processing device to perform a particular function or group of functions. Portions of the computer resources used may be accessed via a network. The computer executable instructions may be, for example, binary, intermediate format instructions such as assembly language, firmware, source code, etc.

Devices implementing the methods and systems described herein may include hardware, software, firmware, middleware, microcode, hardware description language, or any combination thereof, and may take any of a variety of form factors. When implemented in software, firmware, middleware, or microcode, the code or code segments (e.g., computer program products) used to perform the necessary tasks may be stored in a computer-readable or machine-readable medium. The program code may be executed by a processor, which may include one or more processors, such as but not limited to one or more digital signal processors (DSPs), general-purpose microprocessors, application-specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other equivalent integrated or discrete logic circuit systems. Such processors may be configured to implement any of the techniques described in this disclosure. The processor may be a microprocessor, a conventional processor, a controller, a microcontroller, a state machine, or the like. The processor may also be implemented as a combination of computing elements (e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a DSP core, or any other such configuration). Thus, the term "processor" as used herein may refer to any of the aforementioned structures, any combination of the aforementioned structures, or any other structure or device suitable for implementing the techniques described herein. The functionality described herein may also be implemented in a peripheral device or add-in card. As a further example, such functionality may also be implemented in different chips or different processes executed in a single device on a circuit board.

In the foregoing description, various aspects of the disclosure are described with reference to specific examples of the disclosure, but those skilled in the art will recognize that the disclosure is not limited thereto. Therefore, although illustrative examples of the disclosure have been described in detail herein, it should be understood that the inventive concepts may be implemented and employed in different ways, and the appended patent claims are intended to be interpreted as including such variations. The various features and aspects of the above disclosure may be used alone or in any combination. In addition, without departing from the broader spirit and scope of the disclosure, the examples may be used in any number of environments and applications beyond those described herein. Therefore, the disclosure and the figures should be regarded as illustrative rather than restrictive.

The various exemplary logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein may be implemented as electronic hardware, computer software, firmware, or a combination thereof. In order to clearly illustrate this interchangeability of hardware and software, various exemplary components, blocks, modules, circuits, and steps have been generally described above according to their functions. Whether such functions are implemented as hardware or software depends on the specific application and the design constraints on the entire system. A skilled technician may implement the described functions in different ways for each specific application, but such implementation details decisions should not be interpreted as causing a departure from the scope of this application.

Unless otherwise specifically stated, it should be understood that discussions throughout this specification utilizing terms such as "processing," "computing," "calculating," "determining," and "identifying" or similar terms refer to actions or processes of a computing device (e.g., one or more computers or similar electronic computing devices) that manipulate or transform data represented as physical electronic or magnetic quantities within a computing platform's memory, registers, or other information storage devices, transmission devices, or media devices. As used herein, "suitable for" or "configured to" is intended to be an open and inclusive language, and does not exclude devices that are suitable for or configured to perform additional tasks or steps. Additionally, the use of "based on" is meant to be open and inclusive in that a process, step, calculation, or other action "based on" one or more referenced conditions or values may actually be based on additional conditions or values beyond those referenced conditions or values. The headings, lists, and numbers included herein are for ease of explanation only and are not meant to be limiting.

The foregoing detailed description of the technology has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the technology to the precise form disclosed. Many modifications and variations are possible in light of the above teachings. The embodiments described were selected in order to best illustrate the principles of the technology, its practical applications, and to enable others skilled in the art to utilize the technology in a variety of embodiments and with various modifications as are suitable for the particular use contemplated. It is intended that the scope of the technology be defined by the scope of the patent application.

104, 104-1, 104-2, 140-3, 104-n, 600: computing device 108: central processing unit (CPU) 112, 620: memory 116: input/output interface 120: network interface 124: data processor 128, 208: network 132: device 136: model selector 140: feature extractor 144: training data 148, 408: large machine learning model 152, 156, 412, 416, 420, 424: small machine learning model 204: client device 212: content delivery network 404: input sample 428, 432, 436, 440, 444: Model output 452: Benchmark selection 456: Model selection 504, 508, 512, 516, 520, 524: Blocks 602: Cache memory 604: Processor 606: Connectors 608: Storage devices 610: Service 1 612: Service 2 614: Service 3 616: Random access memory (RAM) 618: Read-only memory (ROM) 622: Input devices 624: Output devices 626: Communication interfaces

FIG. 1 illustrates a block diagram of an exemplary system for selecting a machine learning model configured to process disparate data sets according to aspects of the present disclosure. FIG. 2 illustrates an exemplary three-dimensional content representation system according to aspects of the present disclosure. FIG. 3 illustrates a block diagram of an exemplary distributed data processing network according to aspects of the present disclosure. FIG. 4 illustrates a block diagram of an exemplary model selection process for selecting a small machine learning model for processing a data set according to aspects of the present disclosure. FIG. 5 illustrates a flow chart of an exemplary process for model selection for monocular depth estimation according to aspects of the present disclosure. FIG. 6 illustrates an exemplary computing device architecture of an exemplary computing device that can implement various techniques described herein according to aspects of the present disclosure.

104: Computing device 204: Client device 208: Network 212: Content delivery network

Claims (18)

A three-dimensional content representation system includes: a content delivery network, providing representation content including one or more images; a client device, sending a request for the representation content; and a computing device, connected to the content delivery device and the client device, configured to: receive the request from the client device; receive the representation content from the content delivery network according to the request; use a first machine learning model to process one or more images in the representation content to generate a first prediction result; use multiple machine learning models to process the one or more images to generate at least a second prediction result and a third prediction result, wherein the first machine learning model includes more layers than each of the plurality of machine learning models; selecting a second machine learning model from the plurality of machine learning models based on a comparison of the first prediction result with the at least the second prediction result and the third prediction result; and processing the plurality of images using the second machine learning model and transmitting the processing results to the client device, wherein the client device uses the processing results to generate a three-dimensional representation of the representation content. A three-dimensional content representation system as described in claim 1, wherein the computing device includes transmitting the representation content and the depth map obtained by processing the multiple images using the second machine learning model to the client device, and the client device includes generating a three-dimensional representation of the representation content using the representation content and the depth map. A three-dimensional content representation system as described in claim 1, wherein the computing device includes transmitting a depth map obtained by processing the multiple images using the second machine learning model to the client device, and the client device includes receiving the representation content from the content delivery network, and using the representation content and the depth map to generate a three-dimensional representation of the representation content. A three-dimensional content representation system as described in claim 1, wherein the computing device includes a depth map obtained by processing the plurality of images using the second machine learning model, and generates a three-dimensional representation of the representation content using the representation content and the depth map, and transmits or streams the three-dimensional representation of the representation content to the client device for display. A three-dimensional content representation system as described in claim 1, wherein the computing device comprises: generating a first accuracy value of the first prediction result, a second accuracy value of the second prediction result, and a third accuracy value of the third prediction result; and comparing the first accuracy value, the second accuracy value, and the third accuracy value, wherein the second machine learning model is selected based on the second accuracy value being higher than the first accuracy value and the third accuracy value. A three-dimensional content representation system as described in claim 1, wherein the computing device comprises: Using a loss function to generate a first value of the first prediction result, a second value of the second prediction result, and a third value of the third prediction result; and comparing the first value, the second value, and the third value, wherein the second machine learning model is selected based on the second value being higher than the first value and the third value. A three-dimensional content representation method is applicable to a three-dimensional content representation system including a content delivery network, a client device and a computing device, the method comprising the following steps: the computing device receives a request for a representation content from the client device, the representation content including one or more images; the computing device receives the representation content from the content delivery network according to the request; the computing device processes one or more images in the representation content using a first machine learning model to generate a first prediction result; the computing device processes the one or more images using multiple machine learning models , to generate at least a second prediction result and a third prediction result, wherein the first machine learning model includes more layers than each of the plurality of machine learning models; the computing device selects a second machine learning model from the plurality of machine learning models based on a comparison of the first prediction result with the at least the second prediction result and the third prediction result; the computing device processes the one or more images using the second machine learning model and transmits the processing result to the client device; and the client device generates a three-dimensional representation of the representation content using the processing result. The method as described in claim 7 further includes: the computing device transmits the representation content and the depth map obtained by processing the one or more images using the second machine learning model to the client device; and the client device uses the representation content and the depth map to generate a three-dimensional representation of the representation content. The method of claim 7 further includes: the computing device transmits a depth map obtained by processing the one or more images using the second machine learning model to the client device; and the client device receives the representation content from the content delivery network and generates a three-dimensional representation of the representation content using the representation content and the depth map. The method as claimed in claim 7 further includes: the computing device uses the second machine learning model to process the one or more images to obtain a depth map, and uses the representation content and the depth map to generate a three-dimensional representation of the representation content, and transmits or streams the three-dimensional representation of the representation content to the client device for display. The method of claim 7, wherein selecting the second machine learning model from the plurality of machine learning models comprises: generating a first accuracy value for the first prediction result, a second accuracy value for the second prediction result, and a third accuracy value for the third prediction result; and comparing the first accuracy value, the second accuracy value, and the third accuracy value, wherein the second machine learning model is selected based on the second accuracy value being higher than the first accuracy value and the third accuracy value. The method of claim 7, wherein selecting the second machine learning model comprises: using a loss function to generate a first value of the first prediction result, a second value of the second prediction result, and a third value of the third prediction result; and comparing the first value, the second value, and the third value, wherein the second machine learning model is selected based on the second value being higher than the first value and the third value. A three-dimensional content computing device comprises: a communication interface for communicating with a client device and a content delivery network; a storage device for storing instructions for execution by a processor; and a processor, coupled to the communication interface and the storage device, configured to access and execute the instructions stored in the storage device to: receive a request for a representation content from the client device through the communication interface, the representation content comprising one or more images; receive the representation content from the content delivery network through the communication interface according to the request; process one or more images in the representation content using a first machine learning model to generate a first prediction result; Result; Using multiple machine learning models to process the one or more images to generate at least a second prediction result and a third prediction result, wherein the first machine learning model includes more layers than each of the multiple machine learning models; selecting a second machine learning model from the multiple machine learning models based on a comparison of the first prediction result with the at least the second prediction result and the third prediction result; and using the second machine learning model to process the one or more images and transmitting the processing results to the client device, so that the client device uses the processing results to generate a three-dimensional representation of the representation content. A three-dimensional content computing device as described in claim 13, wherein the processor includes transmitting the representation content and the depth map obtained by processing the one or more images using the second machine learning model to the client device through the communication interface, so that the client device uses the representation content and the depth map to generate a three-dimensional representation of the representation content. A three-dimensional content computing device as described in claim 13, wherein the processor includes transmitting a depth map obtained by processing the one or more images using the second machine learning model to the client device through the communication interface, so that the client device receives the representation content from the content delivery network and generates a three-dimensional representation of the representation content using the representation content and the depth map. A three-dimensional content computing device as described in claim 13, wherein the processor includes a depth map obtained by processing the one or more images using the second machine learning model, and generates a three-dimensional representation of the representation content using the representation content and the depth map, and transmits or streams the three-dimensional representation of the representation content to the client device for display via the communication interface. A three-dimensional content computing device as described in claim 13, wherein the processor includes: generating a first accuracy value of the first prediction result, a second accuracy value of the second prediction result, and a third accuracy value of the third prediction result; and comparing the first accuracy value, the second accuracy value, and the third accuracy value, wherein the second machine learning model is selected based on the second accuracy value being higher than the first accuracy value and the third accuracy value. A three-dimensional content computing device as described in claim 13, wherein the processor includes: using a loss function to generate a first value of the first prediction result, a second value of the second prediction result, and a third value of the third prediction result; and comparing the first value, the second value, and the third value, wherein the second machine learning model is selected based on the second value being higher than the first value and the third value.

Images