CN111126179A - Information acquisition method and device, storage medium and electronic device - Google Patents

Abstract

The present application provides an information acquisition method and device, a storage medium and an electronic device, wherein the method includes: extracting key frames from a video to be detected; detecting a first costume of a first costume object in a frame image of the detected key frame region; extract the clothing features of the first clothing object from the first clothing region, wherein the clothing features of the first clothing object include at least one of the following: color information of the first clothing object, posture information of the first clothing object; In the case where a second clothing region in which the clothing features of the included clothing objects and the clothing features of the first clothing object are determined in the multiple reference clothing regions, acquire the target clothing corresponding to the second clothing objects contained in the second clothing region information, wherein each reference clothing area contains at least one clothing object; the region tracking is performed on the first clothing region in the video frame sequence of the video to be detected, and the appearance information of the first clothing object in the video to be detected is determined.

Description

Information acquisition method and device, storage medium and electronic device

technical field

The present application relates to the field of computers, and in particular, to a method and device for acquiring information, a storage medium and an electronic device.

Background technique

At present, when clothing appears in a video program (for example, a variety show), the product information of the same clothing product can be added to provide users with a convenient entrance for purchasing clothing products to meet users' needs for purchasing the same or similar clothing products.

In order to obtain the information of the clothing products included in the video program, it is necessary to identify the clothing styles appearing in the video. The clothing identification method in the related art usually uses clothing feature point information, such as cuff, neckline, and hem position information, to identify clothing styles. However, the above-mentioned clothing recognition methods have limited ability to express clothing features far away from feature points, and are prone to overfitting during training. The algorithm performance in actual scenarios is limited, resulting in the problem of low clothing recognition accuracy.

SUMMARY OF THE INVENTION

The embodiments of the present application provide an information acquisition method and device, a storage medium, and an electronic device, to at least solve the problem of low clothing recognition accuracy due to weak clothing expression ability in clothing identification methods in the related art.

According to an aspect of the embodiments of the present application, a method for acquiring information is provided, which includes: extracting key frames from a video to be detected; detecting a first clothing area including a first clothing object in a frame image of the key frame; The clothing features of the first clothing object are extracted from the clothing area, wherein the clothing features of the first clothing object include at least one of the following: color information of the first clothing object, posture information of the first clothing object; In the case where it is determined that the clothing feature of the clothing object contained in the second clothing area matches the clothing feature of the first clothing object, the target clothing information corresponding to the second clothing object contained in the second clothing area is obtained, wherein each Each reference clothing region contains at least one clothing object; the region tracking is performed on the first clothing region in the video frame sequence of the video to be detected, and the appearance information of the first clothing object in the video to be detected is determined.

According to another aspect of the embodiments of the present application, an apparatus for acquiring information is provided, including: a first extraction unit, configured to extract key frames from a video to be detected; the first clothing area of the first clothing object; the second extraction unit is configured to extract the clothing features of the first clothing object from the first clothing area, wherein the clothing features of the first clothing object include at least one of the following: the first clothing The color information of the object, the posture information of the first clothing object; the first acquisition unit is used to determine the second clothing whose clothing features of the included clothing objects match the clothing features of the first clothing object from a plurality of reference clothing regions In the case of an Area tracking is performed in the video frame sequence of the video to be detected, and the appearance information of the first clothing object in the video to be detected is determined.

Optionally, the first extraction unit includes: a first extraction module, used to extract key frames from the video to be detected according to target intervals; or, a second extraction module, used to extract and shot from the shots included in the video to be detected. corresponding keyframes.

Optionally, the second extraction unit includes: an input module for inputting the first clothing region into the first feature extraction model to obtain clothing features of the first clothing object output by the first feature extraction model, wherein the first feature The extraction model is obtained by training the first initial model by using the first training sample, and the first training sample is an image marked with the first clothing feature of the first training clothing object.

Optionally, the above-mentioned device further includes: a second obtaining unit, configured to obtain the first clothing object before inputting the first clothing region into the first feature extraction model to obtain the clothing features of the first clothing object output by the first feature extraction model. A first clothing feature of a training clothing object, wherein the first clothing feature includes at least one of the following: first color information of the first training clothing object and first posture information of the first training clothing object; a first training unit, using Using the first training sample to train the first initial model to obtain a first feature extraction model, wherein the similarity between the second clothing feature and the first clothing feature extracted by the first feature extraction model from the first training sample is greater than or equal to the first threshold, wherein the second clothing feature includes at least one of the following: second color information of the first training clothing object and second posture information of the first training clothing object.

Optionally, the second obtaining unit includes at least one of the following: an obtaining module, configured to perform histogram calculation and clustering calculation on the first training sample, and obtain first color information; The first position information of the feature points of the training clothing object and the visibility information of the feature points are determined to determine the first posture information.

Optionally, the above-mentioned device further includes: a third acquisition unit, configured to input the clothing features of the first clothing object into the second feature extraction model after extracting the clothing features of the first clothing object from the first clothing area, and obtain the clothing features of the first clothing object. The target feature of the first clothing object output by the second feature extraction model, wherein the second feature extraction model is obtained by using the second training sample to train the second initial model, and the second training sample is marked with the second The clothing feature of the training clothing object and the image of the same item of the second training clothing object, the same item of the second training object is used to identify the second training clothing object of the same style, and the same item extracted by the second feature extraction model The similarity between the target features of the second training clothing objects is greater than or equal to the second threshold, and the similarity between the target features of the second training clothing objects of different styles extracted by the second feature extraction model is less than the second threshold. The 4th acquisition unit is used to obtain from a plurality of reference clothing areas the candidate clothing area that the target feature matches the target feature of the first clothing object; The second determination unit is used to determine the clothing feature from the candidate clothing area and the first A second clothing region to which the clothing features of a clothing object are matched.

Optionally, the above-mentioned device further comprises: an input unit, used for inputting the clothing feature of the first clothing object into the second feature extraction model, and before obtaining the target feature of the first clothing object output by the second feature extraction model. The second training sample is input into the first feature extraction model to obtain the clothing features of the second training clothing object output by the first feature extraction model; the second training unit is used for using the clothing features of the second training clothing object and the second training The same information of the object is used to train the second initial model to obtain the second feature extraction model.

Optionally, the above-mentioned device further includes: an adding unit, and the first determining unit includes a second determining module, wherein the second determining module is used for the video frame before the key frame and the video frame located in the video to be detected according to the first clothing area respectively. The video frame after the key frame is subjected to regional detection, and the time period information and position information of the first clothing object appearing in the video to be detected are determined, wherein the appearance information includes the time period information and the position information; the adding unit is used for A clothing area is tracked in the video frame sequence of the video to be detected, and after determining the appearance information of the first clothing object in the video to be detected, control information is added to the video to be detected, wherein the control information is used to control the video to be detected. When the video is played to the time period corresponding to the time period information, the target clothing information is displayed on the position corresponding to the position information in the video to be detected by means of a pop-up window.

According to yet another embodiment of the present invention, a computer-readable storage medium is also provided, and a computer program is stored in the storage medium, wherein the computer program is configured to execute the steps in any one of the above method embodiments when running.

According to yet another embodiment of the present invention, an electronic device is also provided, comprising a memory and a processor, a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any of the above method embodiments .

Through the present invention, the method of clothing recognition combined with color information and/or posture information is used to extract key frames from the video to be detected; the frame image of the detection key frame includes the first clothing area of the first clothing object; The clothing features of the first clothing object are extracted from the region, wherein the clothing features of the first clothing object include at least one of the following: color information of the first clothing object, posture information of the first clothing object; In the case of determining a second clothing area in which the clothing features of the included clothing objects match those of the first clothing object, obtain target clothing information corresponding to the second clothing objects contained in the second clothing area, wherein each At least one clothing object is included in the reference clothing area; regional tracking is performed on the first clothing area in the video frame sequence of the video to be detected, and the appearance information of the first clothing object in the video to be detected is determined. Clothes recognition by gesture information can achieve the purpose of increasing the expressive ability of clothing features to clothing objects, and achieve the technical effect of improving the accuracy of clothing identification, thereby solving the problem of clothing identification in related technologies due to weak clothing expression ability. Identify problems with low accuracy.

Description of drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description serve to explain the principles of the invention.

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. In other words, on the premise of no creative labor, other drawings can also be obtained from these drawings.

1 is a block diagram of a hardware structure of an optional server according to an embodiment of the present application;

2 is a schematic flowchart of an optional information acquisition method according to an embodiment of the present application;

3 is a schematic flowchart of another optional information acquisition method according to an embodiment of the present application;

4 is a schematic flowchart of an optional model training according to an embodiment of the present application;

5 is a schematic flowchart of an optional clothing feature database construction according to an embodiment of the present application;

FIG. 6 is a schematic flowchart of an optional clothing identification in a video according to an embodiment of the present application; and,

FIG. 7 is a structural block diagram of an optional information acquisition apparatus according to an embodiment of the present application.

Detailed ways

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings and in conjunction with embodiments. It should be noted that the embodiments in the present application and the features of the embodiments may be combined with each other in the case of no conflict.

It should be noted that the terms "first", "second" and the like in the description and claims of the present invention and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific sequence or sequence.

According to an aspect of the embodiments of the present application, a method for acquiring information is provided. Optionally, the method can be executed in a server (a server of a video content playing platform), a user terminal or a similar computing device. Taking running on a server as an example, FIG. 1 is a hardware structural block diagram of an optional server according to an embodiment of the present application. As shown in FIG. 1 , the server 10 may include one or more (only one is shown in FIG. 1 ) processors 102 (the processors 102 may include but are not limited to processing devices such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, optionally, the above server may further include a transmission device 106 and an input and output device 108 for communication functions. Those skilled in the art can understand that the structure shown in FIG. 1 is only for illustration, and does not limit the structure of the above server. For example, server 10 may also include more or fewer components than shown in FIG. 1 , or have a different configuration than that shown in FIG. 1 .

The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as computer programs corresponding to the information acquisition methods in the embodiments of the present application. The processor 102 executes the computer programs stored in the memory 104 by running the computer programs. Various functional applications and data processing implement the above method. Memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, memory 104 may further include memory located remotely from processor 102, which may be connected to server 10 through a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

Transmission means 106 are used to receive or transmit data via a network. The specific example of the above-mentioned network may include a wireless network provided by the communication provider of the server 10 . In one example, the transmission device 106 includes a NIC (Network Interface Controller, network adapter), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be an RF (Radio Frequency, radio frequency) module, which is used to communicate with the Internet in a wireless manner.

This embodiment provides a method for acquiring information running on the above server. FIG. 2 is a flowchart of an optional method for acquiring information according to an embodiment of the present application. As shown in FIG. 2 , the process includes: Follow the steps below:

Step S202, extracting key frames from the video to be detected;

Step S204, the frame image of the detection key frame includes the first clothing area of the first clothing object;

Step S206, extracting the clothing feature of the first clothing object from the first clothing area, wherein the clothing feature of the first clothing object includes at least one of the following: color information of the first clothing object, and posture information of the first clothing object;

Step S208, in the case of determining from a plurality of reference apparel regions a second apparel region in which the apparel features of the included apparel objects match the apparel features of the first apparel object, obtain a second apparel region that matches the second apparel region. target clothing information corresponding to the object, wherein each reference clothing area contains at least one clothing object;

Step S210 , performing regional tracking of the first clothing region in the video frame sequence of the video to be detected, and determining the appearance information of the first clothing object in the video to be detected.

Optionally, the execution subject of the above steps may be a server, a user terminal, etc., but is not limited thereto.

Through the present embodiment, the method of performing clothing identification in combination with color information and/or posture information, because the clothing identification is performed in combination with the color information and/or posture information of the clothing object, the purpose of increasing the expressive ability of clothing features to the clothing object can be achieved, The problem of low clothing recognition accuracy caused by weak clothing expression ability in the clothing recognition method in the related art is solved, and the clothing recognition accuracy is improved.

The method for acquiring information in the embodiment of the present application will be described below with reference to FIG. 2 .

In step S202, key frames are extracted from the video to be detected.

For the program video of a video program (for example, a variety show), the user can watch the program video through a client, a web page, or the like.

In order to display the clothing information (for example, clothing commodity information) corresponding to the clothing objects in the program video in the program video, the clothing information corresponding to the clothing objects in the program video can be pre-added before playing the program video, Or, in the process of playing the program video, the clothing information corresponding to the clothing object included in the program video is determined and acquired in real time.

When acquiring clothing information, firstly extract the key frames of the video to be detected (eg, program video), so as to process the frame images of the extracted key frames. There are many ways to extract the key frames of the video to be detected, which may include but are not limited to one of the following: extraction at equal intervals, extraction by shot.

As an optional embodiment, extracting key frames from the video to be detected includes: extracting key frames from the video to be detected according to target intervals; or, extracting key frames corresponding to shots from shots included in the video to be detected.

As an optional implementation manner, key frames may be extracted from the video to be detected according to target intervals. The above target interval can be a target time interval, for example, every 2s (can be set or modified as needed) to extract a video frame as the current key frame, or it can be a predetermined number of video frame intervals, for example, every 50 video frames (Can be set or modified as needed) Extract a video frame as the current key frame. The specific manner of extracting key frames at equal intervals may be set as required, which is not specifically limited in this embodiment.

As another optional implementation manner, key frames corresponding to the shots may be extracted from shots included in the video to be detected. The video to be detected may include multiple shots, wherein, in the video frames within the same shot, the similarity of adjacent video frames may be greater than or equal to the first threshold. For each shot, one or more video frames can be extracted as keyframes. The specific way of determining the shots included in the video to be detected and the way of extracting key frames from each shot can be set as required (for example, based on the similarity of adjacent video frames), which is not specifically limited in this embodiment.

Through this embodiment, by extracting the key frames of the video to be detected according to a specific interval, or extracting the key frames of the video to be detected according to the shot, the rationality of the key frame extraction can be ensured, and the key frame extraction efficiency can be improved.

In step S204, the frame image of the detected key frame includes the first clothing area of the first clothing object.

For the extracted key frames, the frame images of each key frame may be processed in sequence to determine a first clothing area including the first clothing object in the frame images of each key frame. The first clothing area may be a quadrilateral area containing the first clothing object. The method of detecting the first clothing area may be combined with related technologies, for example, it may be a clothing area detection method based on the fashion detect algorithm, which is not specifically limited herein.

For example, key frame calculation can be performed on the video to be processed, and clothing detection can be performed on each key frame to obtain the clothing area of each key frame.

In step S206, the clothing features of the first clothing object are extracted from the first clothing area, wherein the clothing features of the first clothing object include at least one of the following: color information of the first clothing object, posture information of the first clothing object .

For the detected first clothing region, clothing features of the first clothing object may be extracted from the first clothing region. The clothing features are used to represent clothing objects, and the clothing features may include one or more features, and the clothing features of different clothing objects may be the same or different.

The above-mentioned clothing features can include at least one of the following: color information, posture information, and the color information of the clothing object can be the color information of the entire clothing object, which is used to represent the overall color (main tone) of the clothing object, and the attitude information of the clothing object can be used Used to represent the current pose of the clothing object. In addition to color information and/or posture information, clothing features may also include at least one of the following: clothing category information, style information and feature point information, wherein,

(1) The clothing category information is used for the category of clothing objects, for example, short-sleeved T, long-sleeved hoodie, etc.;

(2) The style information is used to represent the style of the clothing object, for example, finer features such as texture, material, and shape;

(3) The feature point information is used to represent the position information of the feature points of the clothing object, and the above-mentioned feature points may include: two feature points each of the cuff, the neckline, the hem, and the waist.

For the posture information, the posture information of the clothing object can be used to represent the current posture of the clothing object, for example, body (front of the clothing object faces the camera), back (the back of the clothing object faces the camera), sideways (the side of the clothing object faces the camera) and so on. If the current posture of the clothing object is a side view, the posture information may also indicate the angle at which the clothing object is turned with respect to the body. For example, the posture information of the clothing object may indicate that the clothing object is turned sideways by 30°.

As an optional embodiment, extracting the clothing feature of the first clothing object from the first clothing region includes: inputting the first clothing region into the first feature extraction model, and obtaining the first Clothing features of the clothing object, wherein the first feature extraction model is obtained by training the first initial model using the first training sample, and the first training sample is an image marked with the first clothing feature of the first training clothing object included .

For the first clothing region, a first feature extraction model may be used to extract clothing features of the first clothing object. The first clothing region can be input into the first feature extraction model to obtain an output result of the first feature extraction model, and the output result can be used as clothing features of the first clothing object included in the first clothing region.

The first feature extraction model may be obtained by using the first training sample to train the first initial model (the first initial feature extraction model), and the first training sample is marked with the first clothing feature of the first training clothing object included. image.

Through this embodiment, by using the trained first feature extraction model to extract clothing features of the first clothing object, the efficiency of clothing feature extraction and the accuracy of clothing feature extraction can be improved.

Before using the first feature extraction model, the first initial model may be trained by using the first training sample to obtain the first feature extraction model.

As an optional embodiment, before inputting the first clothing region into the first feature extraction model and obtaining the clothing features of the first clothing object output by the first feature extraction model, the first training clothing object's first clothing feature may be obtained. A clothing feature, wherein the first clothing feature includes at least one of the following: first color information of the first training clothing object and first posture information of the first training clothing object; using the first training sample to train the first initial model , obtain a first feature extraction model, wherein the similarity between the second clothing feature extracted from the first training sample and the first clothing feature by the first feature extraction model is greater than or equal to the first threshold, wherein the second clothing feature includes At least one of the following: second color information of the first training clothing object and second posture information of the first training clothing object.

For the first training sample, the first clothing feature of the first training clothing object included in the first training sample may be obtained, and the first clothing feature may include at least one of the following: first color information of the first training clothing object; The first pose information of the first training clothing object. The first clothing feature may further include at least one of the following: first clothing category information of the first training clothing object, first style information of the first training clothing object, and first position information of feature points of the first training clothing object.

After acquiring the first clothing feature of the first training clothing object, use the first training sample to train the first initial model to obtain a first feature extraction model. During model training, each first training sample can be sequentially input into the first initial model in an iterative manner to obtain the detection result output by the first initial model (the clothing feature of the first training clothing object). The degree of similarity between the clothing features of the first training clothing object and the first clothing features output by the model, adjust the model parameters of the first initial model to adjust the clothing features of the first training clothing object output by the first initial model and the first training clothing features. A similarity of clothing features, and when the convergence conditions are met (the objective function of the model training is satisfied), it is determined that the training is completed, and a first feature extraction model is obtained.

The similarity between the second clothing feature extracted from the first training sample and the first clothing feature by the first feature extraction model is greater than or equal to the first threshold, and the second clothing feature may include: second color information of the first training clothing object and the second pose information of the first training apparel object. For example, the first threshold may be the similarity between the clothing features of the first training clothing object and the first clothing features extracted from the first training sample by the adjusted first initial model when the objective function requirements are met.

It should be noted that the server that uses the first feature extraction model and the server that trains the first initial model may be the same server, or may be different servers.

Through this embodiment, by using the first training sample to train the first initial model to obtain the first feature extraction model, the ability of the trained first feature extraction model to detect clothing features of the clothing object can be improved.

There are many ways to obtain the first clothing feature of the first training clothing object. The first clothing feature can be obtained by manual labeling, or the first clothing feature can be obtained by machine labeling.

As an optional embodiment, obtaining the first clothing feature of the first training clothing object includes at least one of the following: performing histogram calculation and clustering calculation on the first training sample to obtain first color information; The first position information of the feature points of the training clothing object and the visibility information of the feature points are determined to determine the first posture information.

For the color information of the first training clothing object, a histogram calculation may be performed on the first training sample to obtain a histogram of the first training sample. Then, according to the histogram, the pixels of the first training sample are clustered, and the color information of the target class with the most pixels in the multiple classes obtained after the clustering is used as the color information of the first training clothing object.

For the posture information of the first training clothing object, the first posture information of the first training clothing object is determined according to the marked first position information of the feature points of the first training clothing object and the visibility information of the feature points. The visibility information of feature points is used to indicate whether the feature points are visible.

According to the marked first location information of the feature points of the first training clothing object and the visibility information of the feature points, the target location area of the first training clothing object (for example, the pattern clothing textures such as the chest and the front of the legs may be relatively The image space attention mask of the rich region) is weighted to obtain the processed first training sample, and the first initial model is trained using the processed first training sample.

For example, the first feature extraction model can be a clothing image feature expression model, which can perform histogram calculation and clustering calculation on a relatively large-scale data set including clothing category, style set, and feature point information, and use semi-automatic data collection. The method mines clothing color information (first color information); combines the position and visibility information of feature points to judge the posture information of the current clothing, and pays attention to the image space in the areas where the patterned clothing textures such as the chest and the front of the legs may be richer. The model is strengthened; combined with the clothing category, style, color, feature points and posture information, the clothing image feature expression model is trained.

Through this embodiment, the color information of the first training clothing object is obtained by clustering the histogram of the training image, and the ability of the first color information to represent the color of the clothing object can be improved; according to the position and visibility of the feature points of the clothing object information, and determine the posture information of the clothing object, which can improve the accuracy of determining the posture information of the clothing object.

After the first feature extraction model is obtained, the first clothing region may be input into the first feature extraction model to obtain an output result of the first feature extraction model, and the output result is used as the clothing feature of the first clothing object.

After the clothing features of the first clothing object are obtained, region matching can be performed directly according to the clothing features of the first clothing object, or the same-style discriminating features (target features) of the first clothing object can be obtained first according to the clothing features of the first clothing object. , and then perform regional screening according to the same judging feature, and finally match the first clothing region with the screened region according to the clothing features of the first clothing object.

As an optional embodiment, after the clothing features of the first clothing object are extracted from the first clothing area, the clothing features of the first clothing object may be input into the second feature extraction model, and the output of the second feature extraction model may be obtained. the target features of the first clothing object; obtain candidate clothing regions whose target features match the target features of the first clothing object from a plurality of reference clothing regions; determine the clothing features and the clothing of the first clothing object from the candidate clothing regions The second clothing area for which the feature is matched.

The second feature extraction model can be used to extract the target feature (discriminative feature of the same item) of the first clothing object. The second feature extraction model can be obtained by training the second initial model by using the second training sample, and the second training sample is marked with the clothing features of the second training clothing object and the same type of identification of the second training clothing object. image. The same style identification of the second training object is used to identify the second training clothing object of the same style (to identify whether the two clothing objects are the same style of clothing). The similarity between the target features of the second training clothing objects of the same style extracted by the second feature extraction model is greater than or equal to the second threshold, and the similarity between the target features of the second training clothing objects of different styles extracted by the second feature extraction model The similarity between the target features is less than the second threshold. The above-mentioned second threshold can be set according to an empirical value, and can be adjusted as required.

After the target feature of the first apparel object is obtained, a candidate apparel region whose target feature matches the target feature of the first apparel object is acquired from a plurality of reference apparel regions. For example, the similarity between the target feature of the first clothing object and the target feature of the reference clothing object contained in each reference clothing area may be calculated, and the difference between the target feature of the reference clothing object contained and the target feature of the first clothing object may be calculated. The reference clothing region whose similarity is greater than or equal to the third threshold is used as a candidate clothing region.

After the candidate clothing regions are obtained, a second clothing region whose clothing features match those of the first clothing object may be determined from the candidate clothing regions. For example, the similarity between the clothing features of the clothing objects contained in the candidate clothing area and the clothing features of the first clothing object can be calculated, and the similarity between the clothing features of the reference clothing objects contained and the clothing features of the first clothing object can be calculated. The candidate clothing area whose similarity is greater than or equal to the fourth threshold is used as the second clothing area.

For example, the image feature expression model (the first feature extraction model) of the clothing image obtained by training can be used to perform image feature calculation on the detected clothing area to obtain the image expression feature of the clothing area (the clothing feature of the first clothing object). The obtained same-style clothing discrimination model (second feature extraction model) is combined with the above-generated image expression features to calculate the same-style discrimination feature (target feature) of the clothing region.

When performing regional retrieval, you can use the calculated discriminant features of the same style to search in the clothing feature database (including multiple reference clothing regions), and use the images in the library that meet the similarity requirements as the same style recommendation candidate (candidate clothing region). ); use the calculated image expression features to confirm the same item of recommendation candidates for the same item, and identify whether the clothing area is consistent with the candidate items in the database. If they are consistent, the consistent candidate clothing area is used as the same item in the video. Apparel Area (Second Apparel Area).

For multiple reference clothing regions, the clothing features and target features of the reference clothing objects in each reference clothing region may be predetermined, and the determined clothing features and target features of the reference clothing objects in each reference clothing region may be stored in the clothing feature database.

For example, an apparel feature database can be pre-built. Perform clothing region detection on clothing images of styles to be promoted, extract image features (clothing features) of clothing region images using the generated clothing image feature expression model, use the generated clothing discrimination model of the same style, and combine the generated image features to extract the clothing features. Distinguishing features of the same style; use the image features and the distinguishing features of the same style as clothing feature vectors to construct a clothing feature database.

Through this embodiment, by first screening the reference clothing regions according to the clothing features of the same style, and then matching the clothing regions according to the clothing features of the clothing objects, the efficiency of clothing region matching can be improved and the occupation of processing resources by clothing region matching can be reduced.

The second initial model may be trained using the second training sample to obtain a second feature recognition model.

As an optional embodiment, before inputting the clothing features of the first clothing object into the second feature extraction model and obtaining the target features of the first clothing object output by the second feature extraction model, inputting the second training sample To the first feature extraction model, obtain the clothing feature of the second training clothing object output by the first feature extraction model; use the clothing feature of the second training clothing object and the same information of the second training object to train the second initial model , to obtain the second feature extraction model.

For the second training sample, the first feature extraction model may be used to extract clothing features of the second training clothing object included in the second training sample. Then, the second initial model is trained by using the clothing features of the second training clothing object and the same item information of the second training object to obtain a second feature extraction model.

When training the second feature extraction model, the clothing features of the second training clothing objects of each second training sample can be input into the second initial model in an iterative manner, and the second training clothing objects output by the second initial model can be obtained. target features. According to the information of the same paragraph, adjust the target features of the second training clothing object included in the current second training sample and the target features of the second training clothing object included in the second training sample before the current second training sample, so that the second training clothing object of the same paragraph The similarity of the target features of the training clothing objects is greater than or equal to the second threshold, and the similarity of the target features of the second training clothing objects of different styles is less than the second threshold.

It should be noted that the target feature can be a discriminative feature of the same item, and can be considered as a joint discriminant feature of multiple clothing features.

For example, a small-scale data set of the same style of clothing can be combined with the generated clothing image feature expression model (the first feature extraction model) to train the similarity discrimination model of the same style of clothing. (Second Feature Extraction Model).

Through this embodiment, the training of the second initial model is performed by combining the clothing features extracted by the first feature extraction model, which can improve the extraction efficiency of the second initial model and simplify the model training process (there is no need to additionally perform the first step in the second training sample). 2. Annotation of clothing features of training clothing objects).

In step S208, in the case of determining a second clothing region in which the clothing features of the included clothing objects match the clothing features of the first clothing object from the plurality of reference clothing regions, obtain the first clothing region that matches the clothing features of the second clothing region. Target clothing information corresponding to two clothing objects, wherein each reference clothing area includes at least one clothing object.

After obtaining the second clothing region matching the first clothing region, target clothing information corresponding to the second clothing object included in the second clothing region may be obtained, and the target clothing information may include but is not limited to: Description information, link information or other information associated with the second clothing object.

Obtaining the target clothing information corresponding to the second clothing object contained in the second clothing area may be: according to the item identifier corresponding to the second clothing object, from a database storing clothing object information of each clothing object, extracting the second clothing Object's target apparel information.

In step S210, area tracking is performed for the first clothing region in the video frame sequence of the video to be detected, and the appearance information of the first clothing object in the video to be detected is determined.

After acquiring the target clothing information corresponding to the second clothing object contained in the second clothing region, the region tracking of the first clothing region in the video frame sequence of the video to be detected can be performed to determine that the first clothing object is in the video to be detected. The above-mentioned appearance information is used to indicate that the first clothing object appears in the video to be detected, and may include but not limited to at least one of the following: time period information (time point information), location information.

The time period (target time period) in which the first clothing object appears in the video to be detected may be one time period or multiple time periods. In each video frame within the target time period, the position information of the first clothing object in the video frame may be the coordinate information of the first clothing object in the video frame (for example, the coordinate information represented by the x, y coordinates), or it may be Region information of the first clothing object in the video frame (for example, the left half region, the right half region, and for example, the middle region, the upper left region, the lower left region, the upper right region, and the lower right region).

The location information of the first clothing object appearing in the video to be detected may be the location information of a specific point (eg, a center point) of the first clothing region appearing in the video to be detected. When storing the position information, the position information of the first clothing object in each video frame within the target time period may be saved, or only the change of the position information of the first clothing object in the video frames within the target time period may be saved .

For example, the first clothing object appears in the video frames from the 5th to the 10th s of the video to be detected, wherein, in the 5th to 7th video frames, the coordinates of the position where the first clothing object appears are (x ₁ , y ₁ ) , in the video frame of the 7th-9th s, the position coordinates of the first clothing object appearing are (x ₂ , y ₂ ), and in the video frame of the 9-10s, the position coordinates of the first clothing object appearing are (x _{3 )} , y ₃ ). Then the time period of the first clothing object in the video to be detected is: 5-10s, and the position of the first clothing object in the video to be detected is: 5-7s, (x ₁ , y ₁ ); 7-9s , (x ₂ , y ₂ ) (or, (x ₂ -x ₁ , y ₂ -y ₁ )); 9-10s, (x ₃ , y ₃ ) (or, (x ₃ -x ₁ , y ₃ -y ₁ ), or, (x ₃ -x ₂ , y ₃ -y ₂ )).

As an optional embodiment, performing regional tracking of the first clothing region in the video frame sequence of the video to be detected, and determining the appearance information of the first clothing object in the video to be detected includes: according to the first clothing region to be detected separately Region detection is performed on the video frame before the key frame and the video frame after the key frame in the video to determine the time period information and position information of the first clothing object appearing in the video to be detected, wherein the appearance information includes the time period information and the position information. location information.

After acquiring the target clothing information corresponding to the second clothing object contained in the second clothing area, it is possible to further determine the time point information at which the first clothing object appears in the video to be detected. The method of determining the time point information may be: bidirectionally tracking the first clothing area in the video frame sequence of the video to be detected, and determining the time period information of the first clothing object appearing in the video to be detected.

There are various ways of bidirectional tracking. For example, the video frame before the key frame and the video frame after the key frame in the video to be detected can be detected, and the first clothing area in the video frame sequence of the video to be detected can be determined. The video frame that appears in the first clothing area is obtained, and the time point information of the appearance of the first clothing area is obtained, and then the time period information of the appearance of the first clothing area is determined. For another example, the key frame located before the current key frame and the key frame located after the current key frame in the video to be detected can be subjected to region detection, and the key frame in which the first clothing region appears in the video frame sequence of the video to be detected can be determined to obtain. The time point information of the appearance of the first clothing area is further determined, and the time period information of the appearance of the first clothing area is further determined.

In addition to the time point information, it is also possible to determine the location information of the candidate region in the video frame in which the first clothing region appears, and then determine the location information of the first clothing object to appear in the video to be detected.

After the region tracking of the first clothing area in the video frame sequence of the video to be detected is performed, and the appearance information of the first clothing object in the video to be detected is determined, control information is added to the video to be detected, wherein the control information is used to control When the video to be detected is played to the time period corresponding to the time period information, the target clothing information is displayed on the position corresponding to the position information in the video to be detected by means of a pop-up window.

After obtaining the time point information and position information of the first clothing object appearing in the video to be detected, control information can be added to the video to be detected, wherein the control information is used to control the video to be detected to be played to the same time period When the information corresponds to the time period, the target clothing information is displayed on the position corresponding to the position information in the video to be detected by means of a pop-up window (or other means).

Through this embodiment, by performing regional tracking on the first clothing area in the video to be detected, the time period information and location information of the first clothing object appearing in the video to be detected can be determined, thereby facilitating the addition of clothing information and improving The accuracy of adding clothing information.

The method for obtaining the above information will be described below with reference to optional examples. This method can be run in the video server. The information acquisition method in this example combines color information and human posture constraints, and adopts a phased model tuning strategy to improve the recognition accuracy of the same style of clothing.

As shown in Figure 3, the method for acquiring information in this example may include the following steps:

Step S302, training a clothing image feature expression model.

Combined with the category, style, color, feature points and posture information of the clothing, an expression model for identifying the characteristics of clothing images, that is, a clothing image feature expression model (as shown in FIG. 4 ) is trained.

Step S304, training a discriminant model for the same style of clothing.

Combined with the small-scale data set of the same style of clothing, the similarity discrimination model of the same style of clothing is trained, that is, the discriminant model, the same style of clothing discriminant model (as shown in Figure 4).

Step S306, constructing a clothing feature database.

Perform clothing area detection on each image in the clothing database, use the expression model to extract the image features (clothing features) of the clothing area images, and then use the discriminant model to extract the same type of clothing discriminant feature (target feature), and build a clothing feature database (such as shown in Figure 5).

In step S308, the region of the same style of clothing in the video is identified.

Perform key frame calculation on the video to be processed, perform clothing detection on each key frame, and calculate the image features and the same type discriminant feature for the detected clothing area, and use the same type discriminant feature to search and query in the clothing feature database. The images in the library that are required by the similarity are used as the recommendation candidates of the same style, and the same style is confirmed by using the image expression features to identify whether the clothing area is consistent with the candidate accessories in the database. shown in Figure 6).

Step S310, generating time point information with frame-level precision for the same style of clothing appearing in the video.

Two-way tracking is performed on the identified clothing area in the video sequence, and the time point information of the clothing appearing in the video is obtained. During this time period, the clothing can be prompted and promoted in the form of a video pop-up window.

Through this example, image feature extraction combined with clothing category, style, color, feature points and posture information can improve the expression ability of clothing features; the staged model training method is adopted to improve the model tuning performance on small data sets; The recognition method of multi-level feature combination improves the recognition accuracy of the same item.

From the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is better implementation. Based on this understanding, the technical solutions of the present invention can be embodied in the form of software products in essence or the parts that make contributions to the prior art, and the computer software products are stored in a storage medium (such as ROM/RAM, magnetic disk, CD-ROM), including several instructions to make a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) execute the methods described in the various embodiments of the present invention.

According to another aspect of the embodiments of the present application, an apparatus for acquiring information for implementing the above method for acquiring information is provided. Optionally, the apparatus is used to implement the above-mentioned embodiments and preferred implementations, and the descriptions that have already been described will not be repeated. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, implementations in hardware, or a combination of software and hardware, are also possible and contemplated.

FIG. 7 is a structural block diagram of an optional information acquisition apparatus according to an embodiment of the present application. As shown in FIG. 7 , the apparatus includes:

(1) The first extraction unit 702 is used to extract key frames from the video to be detected;

(2) the detection unit 704, connected with the first extraction unit 702, is used for detecting the first clothing area of the first clothing object in the frame image of the key frame;

(3) The second extraction unit 706, connected to the detection unit 704, is used for extracting the clothing features of the first clothing object from the first clothing area, wherein the clothing features of the first clothing object include at least one of the following: the first clothing The color information of the object, the posture information of the first clothing object;

(4) The first acquiring unit 708, connected with the second extracting unit 706, is used to determine a second clothing region in which the clothing features of the included clothing objects match the clothing features of the first clothing object from a plurality of reference clothing regions In the case of , obtain the target clothing information corresponding to the second clothing object contained in the second clothing area, wherein each reference clothing area contains at least one clothing object;

(5) The first determination unit 710, connected with the first acquisition unit 708, is used to perform regional tracking of the first clothing area in the video frame sequence of the video to be detected, and to determine the appearance information of the first clothing object in the video to be detected .

Optionally, the first extracting unit 702 may be used in step S202 in the foregoing embodiment, the detecting unit 704 may be used in step S204 in the foregoing embodiment, and the second extraction unit 706 may be used in step S206 in the foregoing embodiment, The first obtaining unit 708 may be configured to perform step S208 in the foregoing embodiments, and the first determining unit 710 may be configured to execute step S210 in the foregoing embodiments.

Through the present embodiment, the method of performing clothing identification in combination with color information and/or posture information, because the clothing identification is performed in combination with the color information and/or posture information of the clothing object, the purpose of increasing the expressive ability of clothing features to the clothing object can be achieved, The problem of low clothing recognition accuracy caused by weak clothing expression ability in the clothing recognition method in the related art is solved, and the clothing recognition accuracy is improved.

As an optional embodiment, the first extraction unit 702 includes:

(1) The first extraction module is used to extract key frames from the video to be detected according to the target interval; or,

(2) The second extraction module is used for extracting key frames corresponding to the shots from the shots included in the video to be detected.

As an optional embodiment, the second extraction unit 706 includes:

(1) an input module, used for inputting the first clothing area into the first feature extraction model to obtain the clothing features of the first clothing object output by the first feature extraction model, wherein the first feature extraction model uses the first training The sample is obtained by training the first initial model, and the first training sample is an image marked with the first clothing feature of the included first training clothing object.

As an optional embodiment, the above device also includes:

(1) The second acquisition unit is used to obtain the first clothing object of the first training clothing before inputting the first clothing area into the first feature extraction model to obtain the clothing features of the first clothing object output by the first feature extraction model. A clothing feature, wherein the first clothing feature includes at least one of the following: first color information of the first training clothing object and first posture information of the first training clothing object;

(2) A first training unit, used for training the first initial model by using the first training sample to obtain a first feature extraction model, wherein the first feature extraction model extracts the second clothing feature from the first training sample The similarity with the first clothing feature is greater than or equal to the first threshold, wherein the second clothing feature includes at least one of the following: second color information of the first training clothing object and second posture information of the first training clothing object.

As an optional embodiment, the second obtaining unit includes at least one of the following:

(1) an acquisition module, for performing histogram calculation and clustering calculation on the first training sample to obtain the first color information;

(2) A first determination module, configured to determine the first posture information according to the first position information of the feature points of the marked first training clothing object and the visibility information of the feature points.

As an optional embodiment, the above device also includes:

(1) The third acquisition unit is used to input the clothing features of the first clothing object into the second feature extraction model after extracting the clothing features of the first clothing object from the first clothing area, and obtain the output of the second feature extraction model , the target feature of the first clothing object, wherein the second feature extraction model is obtained by using the second training sample to train the second initial model, and the second training sample is the clothing feature that marks the included second training clothing object and the image of the same paragraph mark of the second training clothing object, the same paragraph mark of the second training object is used to identify the second training clothing object of the same paragraph, the second training clothing object of the same paragraph extracted by the second feature extraction model The similarity between the target features is greater than or equal to the second threshold, and the similarity between the target features of the second training apparel objects of different models extracted by the second feature extraction model is less than the second threshold;

(2) the 4th acquisition unit, is used to obtain the candidate clothing area that target feature matches with the target feature of the first clothing object from a plurality of reference clothing areas;

(3) A second determining unit, configured to determine, from the candidate clothing regions, a second clothing region whose clothing features match those of the first clothing object.

As an optional embodiment, the above device also includes:

(1) an input unit, used for inputting the clothing features of the first clothing object into the second feature extraction model, and before acquiring the target features of the first clothing object output by the second feature extraction model, inputting the second training sample into The first feature extraction model obtains the clothing features of the second training clothing object output by the first feature extraction model;

(2) a second training unit, used for training the second initial model by using the clothing features of the second training clothing object and the same item information of the second training object to obtain a second feature extraction model.

As an optional embodiment, the above-mentioned apparatus further includes: an adding unit, and the first determining unit includes a second determining module, wherein,

(1) The second determination module is used to perform regional detection on the video frame before the key frame and the video frame after the key frame in the video to be detected according to the first clothing area, and determine that the first clothing object is in the video to be detected. Time period information and location information of occurrence, wherein the occurrence information includes time period information and location information;

(2) an adding unit for performing regional tracking of the first clothing area in the video frame sequence of the video to be detected, and after determining the appearance information of the first clothing object in the video to be detected, adding control information in the video to be detected , wherein the control information is used to control when the video to be detected is played to the time period corresponding to the time period information, the target clothing information is displayed on the position corresponding to the position information in the video to be detected by means of a pop-up window.

It should be noted that the above modules can be implemented by software or hardware, and the latter can be implemented in the following ways, but not limited to this: the above modules are all located in the same processor; or, the above modules can be combined in any combination The forms are located in different processors.

According to yet another aspect of the embodiments of the present application, a computer-readable storage medium is provided. Optionally, a computer program is stored in the storage medium, wherein the computer program is configured to execute the steps in any one of the above methods provided in the embodiments of the present application when the computer program is run.

Optionally, in this embodiment, the above-mentioned storage medium may be configured to store a computer program for executing the following steps:

S1, extract key frames from the video to be detected;

S2, the frame image of the detection key frame includes the first clothing area of the first clothing object;

S3, extract the clothing feature of the first clothing object from the first clothing area, wherein the clothing feature of the first clothing object includes at least one of the following: color information of the first clothing object, and posture information of the first clothing object;

S4, in the case of determining from a plurality of reference clothing regions a second clothing region in which the clothing features of the included clothing objects match the clothing features of the first clothing object, obtain the second clothing object that is included in the second clothing region Corresponding target clothing information, wherein each reference clothing area contains at least one clothing object;

S5, perform regional tracking on the first clothing region in the video frame sequence of the video to be detected, and determine the appearance information of the first clothing object in the video to be detected.

Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: U disk, ROM (Read-Only Memory, read-only memory), RAM (Random Access Memory, random access memory), mobile hard disk, magnetic disk Or various media such as optical discs that can store computer programs.

According to yet another aspect of the embodiments of the present application, an electronic device is provided, including: a processor (the memory may be the processor 102 in FIG. 1 ) and a memory (the memory may be the memory 104 in FIG. 1 ), the memory A computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above methods provided in the embodiments of the present application.

Optionally, the above electronic device may further include a transmission device (the transmission device may be the transmission device 106 in FIG. 1 ) and an input/output device (the input/output device may be the input/output device 108 in FIG. 1 ), wherein the The transmission device is connected to the above-mentioned processor, and the input and output device is connected to the above-mentioned processor.

Optionally, in this embodiment, the above-mentioned processor may be configured to execute the following steps through a computer program:

S1, extract key frames from the video to be detected;

S2, the frame image of the detection key frame includes the first clothing area of the first clothing object;

S3, extract the clothing feature of the first clothing object from the first clothing area, wherein the clothing feature of the first clothing object includes at least one of the following: color information of the first clothing object, and posture information of the first clothing object;

S4, in the case of determining from a plurality of reference clothing regions a second clothing region in which the clothing features of the included clothing objects match the clothing features of the first clothing object, obtain the second clothing object that is included in the second clothing region Corresponding target clothing information, wherein each reference clothing area contains at least one clothing object;

S5, perform regional tracking on the first clothing region in the video frame sequence of the video to be detected, and determine the appearance information of the first clothing object in the video to be detected.

Optionally, for optional examples in this embodiment, reference may be made to the examples described in the foregoing embodiments and optional implementation manners, and details are not described herein again in this embodiment.

Obviously, those skilled in the art should understand that the above-mentioned modules or steps of the present invention can be implemented by a general-purpose computing device, and they can be centralized on a single computing device or distributed in a network composed of multiple computing devices Alternatively, they may be implemented in program code executable by a computing device, such that they may be stored in a storage device and executed by the computing device, and in some cases, in a different order than here The steps shown or described are performed either by fabricating them separately into individual integrated circuit modules, or by fabricating multiple modules or steps of them into a single integrated circuit module. As such, the present invention is not limited to any particular combination of hardware and software.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the principle of the present invention shall be included within the protection scope of the present invention.

Claims (11)

1. An information acquisition method, comprising:

extracting key frames from a video to be detected;

detecting a first decoration area containing a first decoration object in a frame image of the key frame;

extracting a clothing feature of the first clothing object from the first clothing region, wherein the clothing feature of the first clothing object comprises at least one of: color information of the first garment object, and posture information of the first garment object;

under the condition that a second clothing area, which is matched with clothing characteristics of the first clothing object, is determined from a plurality of reference clothing areas, target clothing information corresponding to the second clothing object contained in the second clothing area is obtained, wherein each reference clothing area contains at least one clothing object;

and carrying out region tracking on the first decoration region in the video frame sequence of the video to be detected, and determining the appearance information of the first decoration object in the video to be detected.

2. The method of claim 1, wherein extracting the key frames from the video to be detected comprises:

extracting the key frames from the video to be detected according to the target interval; or,

and extracting the key frames corresponding to the shots from the shots contained in the video to be detected.

3. The method of claim 1, wherein extracting the clothing feature of the first clothing object from the first clothing region comprises:

inputting the first clothing region into a first feature extraction model to obtain clothing features of the first clothing object output by the first feature extraction model, wherein the first feature extraction model is obtained by training a first initial model by using a first training sample, and the first training sample is an image marked with the first clothing features of the first training clothing object.

4. The method of claim 3, wherein prior to inputting the first clothing region to the first feature extraction model, resulting in clothing features of the first clothing object output by the first feature extraction model, the method further comprises:

obtaining the first apparel feature of the first training apparel object, wherein the first apparel feature comprises at least one of: first color information of the first training clothing object and first pose information of the first training clothing object;

training the first initial model by using the first training sample to obtain the first feature extraction model, wherein the similarity between a second clothing feature extracted from the first training sample by the first feature extraction model and the first clothing feature is greater than or equal to a first threshold, and the second clothing feature comprises at least one of the following: second color information of the first training clothing object and second pose information of the first training clothing object.

5. The method of claim 4, wherein obtaining the first garment characteristic of the first training garment object comprises at least one of:

performing histogram calculation and clustering calculation on the first training sample to obtain the first color information;

and determining the first posture information according to the marked first position information of the characteristic points of the first training clothing object and the marked visibility information of the characteristic points.

6. The method of claim 3, wherein after extracting the apparel feature of the first apparel object from the first apparel area, the method further comprises:

inputting the clothing characteristics of the first clothing object into a second characteristic extraction model, acquiring the target characteristics of the first clothing object output by the second characteristic extraction model, wherein the second feature extraction model is obtained by training a second initial model by using a second training sample, the second training sample is an image marked with a clothing feature of a second training clothing object and a homogenous identification of the second training clothing object, the homogeneous identification of the second training object is used to identify the second training apparel object in homogeneous, the similarity between the target features of the second training clothes object extracted by the second feature extraction model and in the same style is larger than or equal to a second threshold value, the similarity between the target features of the second training clothes object extracted by the second feature extraction model and in different styles is smaller than the second threshold value;

obtaining candidate clothing regions with target characteristics matched with the target characteristics of the first clothing object from the plurality of reference clothing regions;

determining the second clothing region with clothing characteristics matched with the clothing characteristics of the first clothing object from the candidate clothing regions.

7. The method of claim 6, wherein before inputting the clothing feature of the first clothing object into the second feature extraction model and obtaining the target feature of the first clothing object output by the second feature extraction model, the method further comprises:

inputting the second training sample into the first feature extraction model to obtain clothing features of the second training clothing object output by the first feature extraction model;

and training the second initial model by using the clothing features of the second training clothing object and the same-style information of the second training object to obtain the second feature extraction model.

8. The method according to any one of claims 1 to 7,

performing region tracking on the first service region in a video frame sequence of the video to be detected, and determining the occurrence information of the first service object in the video to be detected includes: respectively performing area detection on a video frame before the key frame and a video frame after the key frame in the video to be detected according to the first service area, and determining time period information and position information of the first service object appearing in the video to be detected, wherein the appearance information comprises the time period information and the position information;

after performing region tracking on the first service region in the video frame sequence of the video to be detected and determining the occurrence information of the first service object in the video to be detected, the method further includes: adding control information into the video to be detected, wherein the control information is used for controlling the target clothing information to be displayed on the position corresponding to the position information in the video to be detected in a pop-up window mode when the video to be detected is played to the time period corresponding to the time period information.

9. An apparatus for acquiring information, comprising:

the first extraction unit is used for extracting key frames from a video to be detected;

a detecting unit, configured to detect a first decoration area containing a first decoration object in a frame image of the key frame;

a second extraction unit, configured to extract a clothing feature of the first clothing object from the first clothing region, where the clothing feature of the first clothing object includes at least one of: color information of the first garment object, and posture information of the first garment object;

a first obtaining unit, configured to, when a second clothing region in which clothing features of clothing objects included in the plurality of reference clothing regions are matched with clothing features of the first clothing object is determined, obtain target clothing information corresponding to the second clothing object included in the second clothing region, where each of the reference clothing regions includes at least one clothing object;

the first determining unit is configured to perform region tracking on the first service region in the video frame sequence of the video to be detected, and determine occurrence information of the first service object in the video to be detected.

10. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to carry out the method of any one of claims 1 to 8 when executed.

11. An electronic device comprising a memory and a processor, characterized in that the memory has stored therein a computer program, the processor being arranged to execute the method of any of claims 1 to 8 by means of the computer program.

Images