CN107958231B - Light field image filtering method, face analysis method and electronic equipment - Google Patents
Light field image filtering method, face analysis method and electronic equipment Download PDFInfo
- Publication number
- CN107958231B CN107958231B CN201711423426.8A CN201711423426A CN107958231B CN 107958231 B CN107958231 B CN 107958231B CN 201711423426 A CN201711423426 A CN 201711423426A CN 107958231 B CN107958231 B CN 107958231B
- Authority
- CN
- China
- Prior art keywords
- light field
- field image
- face
- image
- face picture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a light field image filtering method, a face analysis method and electronic equipment. The light field image filtering method comprises the following steps: acquiring a light field image to be detected; detecting the light field image to be detected by using a face detector to obtain a face image; determining the ambiguity of the face picture by using a trained ambiguity judgment model; and filtering the face picture according to the fuzziness of the face picture. The invention can judge the fuzziness of the light field image, thereby realizing the effective filtration of the light field image.
Description
Technical Field
The invention relates to the technical field of image recognition, in particular to a light field image filtering method, a human face analysis method and electronic equipment.
Background
The light field technology is the latest technology developed in the field of computer photography, on the basis of the traditional digital imaging theory, the light field technology improves the imaging hardware design by researching and innovating an imaging model, realizes the simultaneous recording of the optical fiber propagation position and angle information in the space, and compared with a common aperture camera, the light field camera has the following characteristics:
(1) the light field camera may record light information in a scene prior to digital refocusing of a designated target.
(2) Because the light field camera adopts multi-camera array collection and rendering, images with high resolution can be synthesized on the basis of a plurality of low-resolution images, and dynamic viewpoint adjustment can be carried out.
(3) The light field camera can complete information and carry out the occlusion removal processing of a specified target, and the technology has very important application in the field of public safety.
Based on the characteristics of the light field camera, because the light field technology uses a plurality of camera arrays when image acquisition is carried out, information of each angle of an object can be captured, and therefore the light field technology can carry out difference values on some invisible areas through a camera model, and lost information is compensated. Due to this property of light field technology, light field cameras have an inherent advantage over traditional cameras for occluding objects. However, due to the limitation of the imaging mode of the light field camera, the rendered video stream can only have a clear image at a certain focal plane, and other video frames have different degrees of blur, and objects appearing in such blurred frames may also be detected, and these detected objects will have a large influence on the subsequent recognition.
Disclosure of Invention
In view of the above, it is necessary to provide a light field image filtering method, a human face analysis method and an electronic device, which can determine the blur degree of a light field image, and further implement effective filtering of the light field image, so as to solve the problem that the blur image in the light field image cannot be effectively filtered, increase the accuracy of a subsequent recognition task, and reduce the system overhead of the subsequent recognition task.
A light field image filtering method, the method comprising:
acquiring a light field image to be detected;
detecting the light field image to be detected by using a face detector to obtain a face image;
determining the ambiguity of the face picture by using a trained ambiguity judgment model;
and filtering the face picture according to the fuzziness of the face picture.
According to a preferred embodiment of the present invention, the acquiring the light field image to be detected includes:
and acquiring a light field image, performing digital refocusing of appointed depth of field on the light field image to obtain a digital refocused light field image, and determining the digital refocused light field image as the light field image to be detected.
According to a preferred embodiment of the present invention, before the digital refocusing of the light field image for the specified depth of field, the method further comprises:
compressing the light field image.
According to a preferred embodiment of the present invention, said compressing said light field image comprises:
compressing the light field image by using a vector quantization algorithm to obtain a quantized compressed light field image;
and processing the quantized compressed light field image by utilizing an entropy coding algorithm to obtain a compressed light field image.
According to a preferred embodiment of the present invention, training the face detector comprises:
acquiring a first training sample by utilizing a web crawler technology, wherein the first training sample comprises positive sample data representing a face picture and negative sample data representing a non-face picture;
and training the face detector by adopting a neural network algorithm according to the first training sample.
According to a preferred embodiment of the present invention, training the ambiguity judging model comprises:
acquiring an acquired light field image;
detecting the collected light field image by using the trained face detector to obtain a light field image with a face;
determining the light field image with the face as a second training sample, wherein the second training sample comprises positive sample data representing a clear picture and negative sample data representing a fuzzy picture;
and taking the second training sample as input data of the face detector, and training the ambiguity judgment model by combining a neural network algorithm.
A method of face analysis, the method comprising:
acquiring a light field image to be detected;
filtering the light field image to be detected by using the light field image filtering method to obtain a reserved human face image;
analyzing and processing the reserved human face picture corresponding to the appointed scene to obtain an analysis result;
and executing the operation corresponding to the specified scene according to the analysis result.
According to a preferred embodiment of the present invention, the analyzing the reserved face picture corresponding to the designated scene to obtain an analysis result, and the executing the operation corresponding to the designated scene according to the analysis result includes:
identifying an image with a target person from the reserved face picture;
and sending the identified image with the target person to at least one terminal device.
According to a preferred embodiment of the present invention, the analyzing the reserved face picture corresponding to the designated scene to obtain an analysis result, and performing the operation corresponding to the designated scene according to the analysis result further includes:
when the face picture is shot by a camera device of a specified vehicle, judging whether a pedestrian exists in the reserved face picture;
and when a pedestrian exists in the face picture, controlling the designated vehicle to brake.
A light field image filtering device, the device comprising:
the acquisition unit is used for acquiring a light field image to be detected;
the detection unit is used for detecting the light field image to be detected by using a face detector to obtain a face image;
the determining unit is used for determining the fuzziness of the face picture by utilizing the trained fuzziness judging model;
and the filtering unit is used for filtering the face picture according to the fuzziness of the face picture.
According to a preferred embodiment of the present invention, the acquiring unit acquiring the light field image to be detected includes:
and acquiring a light field image, performing digital refocusing of appointed depth of field on the light field image to obtain a digital refocused light field image, and determining the digital refocused light field image as the light field image to be detected.
According to a preferred embodiment of the invention, the apparatus further comprises:
and the compression unit is used for compressing the light field image before carrying out digital refocusing of the appointed depth of field on the light field image.
According to a preferred embodiment of the invention, the compression unit is specifically configured to:
compressing the light field image by using a vector quantization algorithm to obtain a quantized compressed light field image;
and processing the quantized compressed light field image by utilizing an entropy coding algorithm to obtain a compressed light field image.
According to a preferred embodiment of the present invention, training the face detector comprises:
the acquisition unit is further used for acquiring a first training sample by utilizing a web crawler technology, wherein the first training sample comprises positive sample data representing a face picture and negative sample data representing a non-face picture;
and the training unit is used for training the face detector by adopting a neural network algorithm according to the first training sample.
According to a preferred embodiment of the present invention, training the ambiguity judging model comprises:
the acquisition unit is also used for acquiring the acquired light field image;
the detection unit is also used for detecting the acquired light field image by using the trained face detector to obtain a light field image with a face;
the determining unit is further configured to determine the light field image with the face as a second training sample, where the second training sample includes positive sample data representing a sharp picture and negative sample data representing a blurred picture;
and the training unit is also used for taking the second training sample as input data of the face detector and training the ambiguity judgment model by combining a neural network algorithm.
An apparatus for face analysis, the apparatus comprising:
the acquisition module is used for acquiring a light field image to be detected;
the filtering module is used for filtering the light field image to be detected by using the light field image filtering method to obtain a reserved human face image;
the analysis module is used for carrying out analysis processing corresponding to the appointed scene on the reserved human face picture to obtain an analysis result;
and the execution module is used for executing the operation corresponding to the specified scene according to the analysis result.
According to a preferred embodiment of the present invention, the analyzing module performs analysis processing corresponding to a specified scene on the retained face picture to obtain an analysis result, and the executing module executes an operation corresponding to the specified scene according to the analysis result, including:
identifying an image with a target person from the reserved face picture;
and sending the identified image with the target person to at least one terminal device.
According to a preferred embodiment of the present invention, the analyzing module performs analysis processing corresponding to a specified scene on the retained face picture to obtain an analysis result, and the executing module executes an operation corresponding to the specified scene according to the analysis result, further including:
when the face picture is shot by a camera device of a specified vehicle, judging whether a pedestrian exists in the reserved face picture;
and when a pedestrian exists in the face picture, controlling the designated vehicle to brake.
An electronic device, the electronic device comprising:
a memory storing at least one instruction; and
a processor executing instructions stored in the memory to implement the light field image filtering method.
A computer-readable storage medium having stored therein at least one instruction, which is executable by a processor in an electronic device to implement the light-field image filtering method.
According to the technical scheme, the light field image to be detected is obtained; detecting the light field image to be detected by using a face detector to obtain a face image; determining the ambiguity of the face picture by using a trained ambiguity judgment model; and filtering the face picture according to the fuzziness of the face picture. The method can judge the fuzziness of the light field image so as to effectively filter the light field image, not only solves the problem that the fuzzy image in the light field image cannot be effectively filtered, but also increases the accuracy of the subsequent identification task and reduces the system overhead of the subsequent identification task.
Drawings
FIG. 1 is a flow chart of a preferred embodiment of the light field image filtering method of the present invention.
FIG. 2 is a flow chart of a preferred embodiment of the face analysis method of the present invention.
FIG. 3 is a functional block diagram of a preferred embodiment of the light field image filtering apparatus of the present invention.
Fig. 4 is a functional block diagram of a face analysis apparatus according to a preferred embodiment of the present invention.
Fig. 5 is a schematic structural diagram of an electronic device implementing a light field image filtering method according to a preferred embodiment of the present invention.
Description of the main elements
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in detail with reference to the accompanying drawings and specific embodiments.
Fig. 1 is a flow chart of a light field image filtering method according to a preferred embodiment of the present invention. The order of the steps in the flow chart may be changed and some steps may be omitted according to different needs.
The light Field image filtering method is applied to one or more electronic devices, which are devices capable of automatically performing numerical calculation and/or information processing according to preset or stored instructions, and the hardware thereof includes, but is not limited to, a microprocessor, an Application Specific Integrated Circuit (ASIC), a Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), an embedded device, and the like.
The electronic device may be any electronic product capable of performing human-computer interaction with a user, for example, a Personal computer, a tablet computer, a smart phone, a Personal Digital Assistant (PDA), a game machine, an interactive Internet Protocol Television (IPTV), an intelligent wearable device, and the like.
The electronic device may also include a network device and/or a user device. The network device includes, but is not limited to, a single network server, a server group consisting of a plurality of network servers, or a Cloud Computing (Cloud Computing) based Cloud consisting of a large number of hosts or network servers.
The Network where the electronic device is located includes, but is not limited to, the internet, a wide area Network, a metropolitan area Network, a local area Network, a Virtual Private Network (VPN), and the like.
S10, the electronic equipment acquires a light field image to be detected.
In at least one embodiment of the present invention, the electronic device may acquire the light field image to be detected through a light field camera in communication with the electronic device.
Specifically, the electronic device may employ a multi-camera array to acquire the light field image to be detected, for example: the electronic device can acquire image data and the like acquired by the six-way camera array. Therefore, the electronic equipment can fully record the multi-channel data information in the acquisition scene for subsequent use when the digital refocusing is carried out on the specified target.
In at least one embodiment of the present invention, the electronic device acquiring the light field image to be detected includes:
the electronic equipment acquires a light field image, performs digital refocusing of a specified depth of field on the light field image to obtain a digital refocused light field image, and determines the digital refocused light field image as the light field image to be detected.
Therefore, a user of the electronic equipment can set the appointed depth of field according to actual requirements, and digital refocusing of the appointed depth of field is carried out on the light field image to obtain a focusing image required by the user, so that the image can be obtained more specifically, data redundancy caused by unnecessary images is reduced, and the processing efficiency is improved.
In at least one embodiment of the invention, prior to digitally refocusing the light field image for a specified depth of field, the method further comprises:
the electronic device compresses the light field image.
It should be noted that, because the data size of the light field image is large, which is not beneficial to storage, and the storage speed is reduced, before the digital refocusing of the specified depth of field is performed on the light field image, the electronic device compresses the light field image, so as to improve the efficiency of storing and processing the light field image.
In at least one embodiment of the present invention, the electronic device compressing the light field image comprises:
the electronic equipment compresses the light field image by using a Vector Quantization (VQ) algorithm to obtain a quantized compressed light field image, and then processes the quantized compressed light field image by using an entropy coding algorithm to obtain a compressed light field image.
In at least one embodiment of the present invention, the vector quantization algorithm is an important signal compression method, and the basic idea of the vector quantization algorithm is: several scalar data groups are formed into a vector, and then the vector space is subjected to integral quantization, so that the data is compressed, and less information is lost.
In this embodiment, the vector quantization algorithm is adopted because it is efficient, has a large compression ratio, is simple to decode, and has small distortion.
In at least one embodiment of the present invention, the entropy coding algorithm is a coding mode that does not lose any information during the coding process according to the entropy principle.
In particular, the entropy coding algorithm includes, but is not limited to: shannon (Shannon) coding, Huffman (Huffman) coding, arithmetic coding (arithmet) and the like.
In at least one embodiment of the present invention, the electronic device compresses the light field image by using a vector quantization algorithm, and the obtained quantized compressed light field image is lossy data, so that the electronic device processes the quantized compressed light field image by using an entropy coding algorithm to obtain a compressed light field image.
In this way, because the encoding process of the entropy encoding algorithm is an encoding mode without losing any information according to the entropy principle, the information amount is not lost in the encoding process, and the lossless compression of the light field image is realized.
In at least one embodiment of the present invention, the electronic device digitally refocusing the light field image for a specified depth of field comprises:
note LFF (x, y, u, v) is a mathematical description of the light-field image, where LFFor a given amount of radiation of a ray, u, v are focal planes, x, y are image planes, and F denotes the distance between the focal plane and the image plane. Then, the light field image may be expressed as the following formula (1), and at this time, if the F is adjusted to F', the new image may be expressed as the following formula (2).
Wherein α represents a refocusing factor, BαRepresentation fourierThe coefficient of the inner lobe.
Through the process, the electronic equipment can realize digital refocusing of the light field image in the designated depth of field.
And S11, the electronic equipment detects the light field image to be detected by using a human face detector to obtain a human face picture.
In at least one embodiment of the present invention, before the electronic device detects the light field image to be detected by using a face detector to obtain a face picture, the method further includes:
the electronic device trains the face detector.
In at least one embodiment of the invention, the electronic device training the face detector comprises:
the electronic equipment acquires a first training sample by utilizing a web crawler technology, wherein the first training sample comprises positive sample data representing a face picture and negative sample data representing a non-face picture, and the electronic equipment trains the face detector by adopting a neural network algorithm according to the first training sample.
In at least one embodiment of the present invention, on one hand, the face detector is configured to detect the light field image to be detected to obtain a face image, so as to perform further ambiguity determination on the basis of the face image in the following process; on the other hand, the face detector may also be used as a basis for training the ambiguity resolution model.
In at least one embodiment of the present invention, since the amount of data of the light field image obtained by the light field camera communicating with the electronic device is small, the ambiguity judgment model is trained by directly using the clear and blurred face pictures captured by the light field camera as the training samples, which is not enough to reflect the distribution of the training samples, and the classification of the trained ambiguity judgment model is not accurate, therefore, in this embodiment, the electronic device first obtains a large amount of face data as the first training sample by using a web crawler technology, the first training sample includes positive sample data representing the face pictures and negative sample data representing non-face pictures, the electronic device learns the structural features of the face according to the first training sample, and obtains the face detector by training through a neural network algorithm, in this way, the electronic device may then adjust the clear and blurred face pictures captured by the light field camera based on the trained face detector to obtain the ambiguity resolution model, and further implement the classification of the face pictures.
In at least one embodiment of the invention, the electronic device trains the face detector using a neural network algorithm.
Specifically, the electronic device normalizes input face image data to achieve unification of data types and formats, in this embodiment, the input face image data is normalized to 48 × 48, the electronic device inputs the input face image data into 4 Convolutional layers (Convolutional layers) of 3 × 3, maximum pooling layers (maximum pooling) of 3 × 3, and full connected layers (full connected layer), and finally, classification of the input face image data and prediction of a face position in the input face image data are completed through a loss function and a euclidean distance algorithm, respectively.
Specifically, since the electronic device adopts the neural network algorithm to train the face detector, which is already relatively mature in the prior art, the present invention is not described herein again.
And S12, the electronic equipment determines the ambiguity of the face picture by using the trained ambiguity judgment model.
In at least one embodiment of the present invention, before the electronic device determines the ambiguity of the face picture by using the trained ambiguity determination model, the method further includes:
the electronic device trains the ambiguity judging model.
In at least one embodiment of the present invention, the electronic device training the ambiguity judging model includes:
the electronic equipment acquires a collected light field image, the collected light field image is detected by the trained face detector to obtain a light field image with a face, the electronic equipment determines the light field image with the face as a second training sample, the second training sample comprises positive sample data representing a clear picture and negative sample data representing a fuzzy picture, and the electronic equipment takes the second training sample as input data of the face detector and trains the ambiguity judging model by combining a neural network algorithm.
Specifically, the electronic device inputs the second training sample as input data into the face detector, then randomly initializes parameters corresponding to a fully-connected layer of the network of the face detector to any value in 0-1, and measures the loss of the fully-connected layer by using a loss function to complete the classification of the second training sample.
Specifically, since the electronic device adopts the neural network algorithm to train the ambiguity determination model, which is relatively mature in the prior art, the invention is not described herein again.
And S13, the electronic equipment filters the face picture according to the fuzziness of the face picture.
In at least one embodiment of the present invention, the filtering, by the electronic device, the face picture according to the ambiguity of the face picture includes:
and when the fuzziness of the face picture is smaller than the configured fuzziness value, the electronic equipment reserves the face picture.
Therefore, the electronic equipment can perform subsequent analysis processing through the reserved human face picture, so that the subsequent processing data is more accurate. For a specific application, see the following examples.
Or, when the blurring degree of the face picture is greater than or equal to the configured blurring value, the electronic device discards the face picture.
Therefore, the electronic equipment can directly discard the data of the blurred image without use value, so that the storage space is saved, and the phenomenon of inaccurate analysis caused by the interference of the blurred image when the face image is subsequently utilized for analysis processing can be avoided, thereby improving the efficiency of analysis processing.
It should be noted that, the value of the configured fuzzy value is not limited in the present invention, and the configuration may be performed according to actual needs.
In conclusion, the invention can obtain the light field image to be detected; detecting the light field image to be detected by using a face detector to obtain a face image; determining the ambiguity of the face picture by using a trained ambiguity judgment model; and filtering the face picture according to the fuzziness of the face picture. Therefore, the method can judge the fuzziness of the light field image, further realize effective filtration of the light field image, not only solve the problem that the fuzzy image in the light field image cannot be effectively filtered, but also increase the accuracy of the subsequent identification task and reduce the system overhead of the subsequent identification task.
Fig. 2 is a flow chart of a preferred embodiment of the face analysis method of the present invention. The order of the steps in the flow chart may be changed and some steps may be omitted according to different needs.
S20, the electronic equipment acquires a light field image to be detected.
And S21, the electronic equipment filters the light field image to be detected by using the light field image filtering method to obtain a reserved human face image.
And S22, the electronic equipment performs analysis processing corresponding to the appointed scene on the reserved human face picture to obtain an analysis result.
And S23, the electronic equipment executes the operation corresponding to the appointed scene according to the analysis result.
In at least one embodiment of the invention, the electronic device can identify the person in the reserved face picture and perform further processing on the person.
Preferably, the analyzing, by the electronic device, the reserved face picture corresponding to the designated scene to obtain an analysis result, and the executing, according to the analysis result, the operation corresponding to the designated scene includes:
and the electronic equipment identifies the image with the target person from the reserved human face picture and sends the identified image with the target person to at least one terminal device.
Further, when the target person is a lost person, the electronic device acquires the shooting time and the shooting place of the image of the lost person, and sends the image of the lost person and the acquired shooting time and the shooting place of the image of the lost person to a specified user device.
Specifically, the electronic device may record the time when the light field camera takes the image of the lost person, or the light field camera displays the taking time on the image of the lost person when taking the image of the lost person, or the like.
Specifically, the electronic device may record a shooting location at which the light field camera shoots an image of the lost person, and determine the shooting location as a location where the lost person appears.
Therefore, the electronic equipment can timely acquire the time and place of the lost person in the above manner, so as to help related persons (such as family members or police) to find the lost person more quickly.
Further, when the target person is a dangerous person, the electronic device acquires the shooting time and the shooting place of the image of the dangerous person, and sends the image of the dangerous person and the acquired shooting time and the shooting place of the image of the dangerous person to a police server to which the shooting place belongs.
Therefore, the electronic equipment can realize alarm in time and quickly, a shot clear picture and the obtained time and place are used as auxiliary information, and in addition, the information is sent to a police service server to which the shooting place belongs, so that a police officer closest to the shooting place can also confirm the target person quickly and prepare for catching, and the catching efficiency is improved.
In at least one embodiment of the present invention, the electronic device may analyze a face picture taken by a camera of a vehicle to control braking of the vehicle.
Preferably, the analyzing processing corresponding to the specified scene is performed on the reserved human face picture by the electronic device to obtain an analysis result, and the executing the operation corresponding to the specified scene according to the analysis result further includes:
when the face picture is shot by a camera of a designated vehicle, the electronic equipment judges whether a pedestrian exists in the reserved face picture, and controls the designated vehicle to brake when the pedestrian exists in the face picture.
For example: when the electronic equipment passes through a crossroad, the electronic equipment judges whether a pedestrian exists in the identified clear face picture, specifically, the electronic equipment can judge whether the corresponding person walks through the limb action of the corresponding person in the face picture, when the electronic equipment judges that the corresponding person walks, the electronic equipment judges that the pedestrian exists, and the electronic equipment controls the specified vehicle to brake.
Therefore, the electronic equipment can ensure the driving safety of the vehicle in an emergency braking mode and can also play a certain safety protection role in the field of unmanned driving.
Fig. 3 is a functional block diagram of a light field image filtering apparatus according to a preferred embodiment of the present invention. The light field image filtering apparatus 11 includes an acquisition unit 110, a detection unit 111, a determination unit 112, a filtering unit 113, a compression unit 114, and a training unit 115. The module/unit referred to in the present invention refers to a series of computer program segments that can be executed by the processor 13 and that can perform a fixed function, and that are stored in the memory 12. In the present embodiment, the functions of the modules/units will be described in detail in the following embodiments.
The acquisition unit 110 acquires a light field image to be detected.
In at least one embodiment of the present invention, the acquiring unit 110 may acquire the light field image to be detected through a light field camera in communication with the electronic device.
Specifically, the acquiring unit 110 may employ a multi-camera array to acquire the light field image to be detected, for example: the acquisition unit 110 may acquire image data collected by a six-way camera array, and the like. In this way, the obtaining unit 110 can fully record the multi-channel data information in the captured scene for subsequent use in digital refocusing of the designated target.
In at least one embodiment of the present invention, the acquiring unit 110 acquires the light field image to be detected, including:
the acquiring unit 110 acquires a light field image, performs digital refocusing on the light field image with a specified depth of field to obtain a digital refocused light field image, and the acquiring unit 110 determines the digital refocused light field image as the light field image to be detected.
Therefore, a user of the electronic equipment can set the appointed depth of field according to actual requirements, and digital refocusing of the appointed depth of field is carried out on the light field image to obtain a focusing image required by the user, so that the image can be obtained more specifically, data redundancy caused by unnecessary images is reduced, and the processing efficiency is improved.
In at least one embodiment of the invention, prior to digitally refocusing the light field image for a specified depth of field, the method further comprises:
the compression unit 114 compresses the light field image.
It should be noted that, because the data size of the light field image is large, which is not beneficial to storage, and the storage speed is reduced, before the digital refocusing of the specified depth of field is performed on the light field image, the compression unit 114 compresses the light field image, so as to improve the efficiency of storing and processing the light field image.
In at least one embodiment of the present invention, the compressing unit 114 compressing the light field image comprises:
the compression unit 114 compresses the light field image by using a Vector Quantization (VQ) algorithm to obtain a quantized compressed light field image, and then processes the quantized compressed light field image by using an entropy coding algorithm to obtain a compressed light field image.
In at least one embodiment of the present invention, the vector quantization algorithm is an important signal compression method, and the basic idea of the vector quantization algorithm is: several scalar data groups are formed into a vector, and then the vector space is subjected to integral quantization, so that the data is compressed, and less information is lost.
In this embodiment, the vector quantization algorithm is adopted because it is efficient, has a large compression ratio, is simple to decode, and has small distortion.
In at least one embodiment of the present invention, the entropy coding algorithm is a coding mode that does not lose any information during the coding process according to the entropy principle.
In particular, the entropy coding algorithm includes, but is not limited to: shannon (Shannon) coding, Huffman (Huffman) coding, arithmetic coding (arithmet) and the like.
In at least one embodiment of the present invention, the compression unit 114 performs compression processing on the light field image by using a vector quantization algorithm, and the obtained quantized compressed light field image is lossy data, so the compression unit 114 performs processing on the quantized compressed light field image by using an entropy coding algorithm to obtain a compressed light field image.
In this way, because the encoding process of the entropy encoding algorithm is an encoding mode without losing any information according to the entropy principle, the information amount is not lost in the encoding process, and the lossless compression of the light field image is realized.
In at least one embodiment of the present invention, the electronic device digitally refocusing the light field image for a specified depth of field comprises:
note LFF (x, y, u, v) is a mathematical description of the light-field image, where LFFor a given amount of radiation, u, v are focal planes, x, y are image planes, and F representsIs the distance between the focal plane and the image plane. Then, the light field image may be expressed as the following formula (1), and at this time, if the F is adjusted to F', the new image may be expressed as the following formula (2).
Wherein α represents a refocusing factor, BαRepresenting fourier coefficients.
Through the process, the electronic equipment can realize digital refocusing of the light field image in the designated depth of field.
The detection unit 111 detects the light field image to be detected by using a face detector to obtain a face image.
In at least one embodiment of the present invention, before the detecting unit 111 detects the light field image to be detected by using a face detector to obtain a face picture, the method further includes:
the training unit 115 trains the face detector.
In at least one embodiment of the present invention, the training unit 115 trains the face detector comprises:
the training unit 115 obtains a first training sample by using a web crawler technology, the first training sample includes positive sample data representing a face picture and negative sample data representing a non-face picture, and the training unit 115 trains the face detector according to the first training sample and by using a neural network algorithm.
In at least one embodiment of the present invention, on one hand, the face detector is configured to detect the light field image to be detected to obtain a face image, so as to perform further ambiguity determination on the basis of the face image in the following process; on the other hand, the face detector may also be used as a basis for training the ambiguity resolution model.
In at least one embodiment of the present invention, since the amount of data of the light field image obtained by the light field camera communicating with the electronic device is small, the ambiguity judgment model is trained by directly using the clear and blurred face pictures captured by the light field camera as the training samples, which is not enough to reflect the distribution of the training samples, and therefore, the ambiguity judgment model is not accurately classified, in this embodiment, the training unit 115 first obtains a large amount of face data as the first training sample by using a web crawler technology, the first training sample includes positive sample data representing the face pictures and negative sample data representing non-face pictures, the training unit 115 learns the structural features of the face according to the first training sample and trains the face detector by using a neural network algorithm, in this way, the training unit 115 may then adjust the clear and blurred face pictures captured by the light field camera based on the trained face detector to obtain the ambiguity resolution model, and further implement the classification of the face pictures.
In at least one embodiment of the present invention, the training unit 115 trains the face detector using a neural network algorithm.
Specifically, the training unit 115 normalizes the input face picture data to realize unification of data types and formats, in this embodiment, the input face picture data is normalized to 48 × 48, the training unit 115 inputs the input face picture data into 43 × 3 Convolutional layers (Convolutional layers), 3 × 3 maximum pooling layers (maximum pooling), and 1 full connected layer (full connected layer), and finally completes classification of the input face picture data and prediction of a face position in the input face picture data through a loss function and a euclidean distance algorithm, respectively.
Specifically, since the training unit 115 adopts a neural network algorithm to train the face detector, which is already relatively mature in the prior art, the present invention is not described herein again.
The determining unit 112 determines the ambiguity of the face picture by using the trained ambiguity judging model.
In at least one embodiment of the present invention, before the determining unit 112 determines the ambiguity of the face picture by using the trained ambiguity judging model, the method further includes:
the training unit 115 trains the ambiguity determination model.
In at least one embodiment of the present invention, the training unit 115 trains the ambiguity judging model includes:
the training unit 115 obtains an acquired light field image, the acquired light field image is detected by the trained face detector to obtain a light field image with a face, the training unit 115 determines the light field image with the face as a second training sample, the second training sample comprises positive sample data representing a sharp picture and negative sample data representing a fuzzy picture, and the training unit 115 takes the second training sample as input data of the face detector and trains the ambiguity judgment model by combining a neural network algorithm.
Specifically, the training unit 115 inputs the second training sample as input data to the face detector, then randomly initializes parameters corresponding to a fully-connected layer of the network of the face detector to any value in 0-1, and measures a loss of the fully-connected layer by using a loss function to complete classification of the second training sample.
Specifically, since the training unit 115 trains the ambiguity determination model by using a neural network algorithm, which is already relatively mature in the prior art, the present invention is not described herein again.
The filtering unit 113 filters the face picture according to the blur degree of the face picture.
In at least one embodiment of the present invention, the filtering unit 113, according to the ambiguity of the face picture, includes:
when the blur degree of the face picture is smaller than the configured blur value, the filtering unit 113 retains the face picture.
In this way, the filtering unit 113 can perform subsequent analysis processing on the retained face picture, so as to make the subsequently processed data more accurate. For a specific application, see the following examples.
Alternatively, when the degree of blur of the face picture is greater than or equal to the configured blur value, the filtering unit 113 discards the face picture.
In this way, the filtering unit 113 can directly discard the data of the blurred image without use value, which not only saves the storage space, but also can avoid the occurrence of inaccurate analysis caused by the interference of the blurred image when the face image is subsequently utilized for analysis, thereby improving the efficiency of analysis.
It should be noted that, the value of the configured fuzzy value is not limited in the present invention, and the configuration may be performed according to actual needs.
In conclusion, the invention can obtain the light field image to be detected; detecting the light field image to be detected by using a face detector to obtain a face image; determining the ambiguity of the face picture by using a trained ambiguity judgment model; and filtering the face picture according to the fuzziness of the face picture. Therefore, the method can judge the fuzziness of the light field image, further realize effective filtration of the light field image, not only solve the problem that the fuzzy image in the light field image cannot be effectively filtered, but also increase the accuracy of the subsequent identification task and reduce the system overhead of the subsequent identification task.
Fig. 4 is a functional block diagram of a face analysis apparatus according to a preferred embodiment of the present invention. The face analysis device 14 includes an obtaining module 141, a filtering module 142, an analyzing module 143, and an executing module 144. The module/unit referred to in the present invention refers to a series of computer program segments that can be executed by the processor of the face analysis device 14 and can perform a fixed function, and is stored in the memory of the face analysis device 14. In the present embodiment, the functions of the modules/units will be described in detail in the following embodiments.
The acquisition module 141 acquires a light field image to be detected.
The filtering module 142 filters the light field image to be detected by using the light field image filtering method to obtain a reserved human face image.
The analysis module 143 performs analysis processing corresponding to the designated scene on the retained face picture to obtain an analysis result.
The execution module 144 executes an operation corresponding to the designated scene according to the analysis result.
In at least one embodiment of the invention, the electronic device can identify the person in the reserved face picture and perform further processing on the person.
Preferably, the analyzing module 143 performs analysis processing corresponding to a specified scene on the retained face picture to obtain an analysis result, and the executing module 144 executes an operation corresponding to the specified scene according to the analysis result, where the executing operation includes:
the analysis module 143 identifies an image with a target person from the retained face picture, and the execution module 144 transmits the identified image with the target person to at least one terminal device.
Further, when the target person is a lost person, the analysis module 143 obtains a shooting time and a shooting location of the image of the lost person, and the execution module 144 sends the image of the lost person and the obtained shooting time and shooting location of the image of the lost person to a designated user device.
Specifically, the electronic device may record the time when the light field camera takes the image of the lost person, or the light field camera displays the taking time on the image of the lost person when taking the image of the lost person, or the like.
Specifically, the electronic device may record a shooting location at which the light field camera shoots an image of the lost person, and determine the shooting location as a location where the lost person appears.
Therefore, the electronic equipment can timely acquire the time and place of the lost person in the above manner, so as to help related persons (such as family members or police) to find the lost person more quickly.
Further, when the target person is a dangerous person, the analysis module 143 obtains a photographing time and a photographing place of the image of the dangerous person, and the execution module 144 sends the image of the dangerous person and the obtained photographing time and the photographing place of the image of the dangerous person to a police server to which the photographing place belongs.
Therefore, the electronic equipment can realize alarm in time and quickly, a shot clear picture and the obtained time and place are used as auxiliary information, and in addition, the information is sent to a police service server to which the shooting place belongs, so that a police officer closest to the shooting place can also confirm the target person quickly and prepare for catching, and the catching efficiency is improved.
In at least one embodiment of the present invention, the electronic device may analyze a face picture taken by a camera of a vehicle to control braking of the vehicle.
Preferably, the analyzing module 143 performs analysis processing corresponding to a specified scene on the retained face picture to obtain an analysis result, and the executing module 144 executes an operation corresponding to the specified scene according to the analysis result, further includes:
when the face picture is taken by a camera of a designated vehicle, the analysis module 143 determines whether there is a pedestrian in the retained face picture, and when there is a pedestrian in the face picture, the execution module 144 controls the designated vehicle to brake.
For example: when the electronic device passes through an intersection, the analysis module 143 determines whether there is a pedestrian from the identified clear face picture, specifically, the analysis module 143 may determine whether the corresponding person is walking through the body motion of the person corresponding to the face picture, and when the analysis module 143 determines that the corresponding person is walking, the analysis module 143 determines that there is a pedestrian, and the execution module 144 controls the designated vehicle to brake.
Therefore, the electronic equipment can ensure the driving safety of the vehicle in an emergency braking mode and can also play a certain safety protection role in the field of unmanned driving.
Fig. 5 is a schematic structural diagram of an electronic device implementing a light field image filtering method according to a preferred embodiment of the present invention.
The electronic device 1 is a device capable of automatically performing numerical calculation and/or information processing according to a preset or stored instruction, and its hardware includes, but is not limited to, a microprocessor, an Application Specific Integrated Circuit (ASIC), a Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), an embedded device, and the like.
The electronic device 1 may also be, but not limited to, any electronic product that can perform human-computer interaction with a user through a keyboard, a mouse, a remote controller, a touch panel, or a voice control device, for example, a Personal computer, a tablet computer, a smart phone, a Personal Digital Assistant (PDA), a game machine, an Internet Protocol Television (IPTV), an intelligent wearable device, and the like.
The electronic device 1 may also be a desktop computer, a notebook, a palm computer, a cloud server, or other computing devices.
The Network where the electronic device 1 is located includes, but is not limited to, the internet, a wide area Network, a metropolitan area Network, a local area Network, a Virtual Private Network (VPN), and the like.
In one embodiment of the present invention, the electronic device 1 includes, but is not limited to, a memory 12, a processor 13, and a computer program, such as a light field image filtering program, stored in the memory 12 and executable on the processor 13.
It will be appreciated by a person skilled in the art that the schematic diagram is only an example of the electronic device 1 and does not constitute a limitation of the electronic device 1, and that it may comprise more or less components than shown, or some components may be combined, or different components, e.g. the electronic device 1 may further comprise an input output device, a network access device, a bus, etc.
The Processor 13 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. The processor 13 is an operation core and a control center of the electronic device 1, and is connected to each part of the whole electronic device 1 by various interfaces and lines, and executes an operating system of the electronic device 1 and various installed application programs, program codes, and the like.
The processor 13 executes an operating system of the electronic device 1 and various installed application programs. The processor 13 executes the application program to implement the steps in the above-described respective light field image filtering method embodiments, such as the steps S10, S11, S12, S13 shown in fig. 1.
Alternatively, the processor 13, when executing the computer program, implements the functions of the modules/units in the above device embodiments, for example: acquiring a light field image to be detected; detecting the light field image to be detected by using a face detector to obtain a face image; determining the ambiguity of the face picture by using a trained ambiguity judgment model; and filtering the face picture according to the fuzziness of the face picture.
Illustratively, the computer program may be divided into one or more modules/units, which are stored in the memory 12 and executed by the processor 13 to accomplish the present invention. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution process of the computer program in the electronic device 1. For example, the computer program may be divided into an acquisition unit 110, a detection unit 111, a determination unit 112, a filtering unit 113, a compression unit 114 and a training unit 115.
The memory 12 can be used for storing the computer programs and/or modules, and the processor 13 implements various functions of the electronic device 1 by running or executing the computer programs and/or modules stored in the memory 12 and calling data stored in the memory 12. The memory 12 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data (such as audio data, a phonebook, etc.) created according to the use of the cellular phone, and the like. Further, the memory 12 may include high speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.
The memory 12 may be an external memory and/or an internal memory of the electronic device 1. Further, the Memory 12 may be a circuit having a Memory function without any physical form In the integrated circuit, such as a RAM (Random-Access Memory), a FIFO (First In First Out), and the like. Alternatively, the memory 12 may be a memory in a physical form, such as a memory stick, a TF Card (Trans-flash Card), or the like.
The integrated modules/units of the electronic device 1 may be stored in a computer-readable storage medium if they are implemented in the form of software functional units and sold or used as separate products. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments may be implemented.
Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable medium may contain content that is subject to appropriate increase or decrease as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media does not include electrical carrier signals and telecommunications signals as is required by legislation and patent practice.
With reference to fig. 1, the memory 12 in the electronic device 1 stores a plurality of instructions to implement a light field image filtering method, and the processor 13 can execute the plurality of instructions to implement: acquiring a light field image to be detected; detecting the light field image to be detected by using a face detector to obtain a face image; determining the ambiguity of the face picture by using a trained ambiguity judgment model; and filtering the face picture according to the fuzziness of the face picture.
According to a preferred embodiment of the present invention, the processor 13 further executes a plurality of instructions including:
and acquiring a light field image, performing digital refocusing of appointed depth of field on the light field image to obtain a digital refocused light field image, and determining the digital refocused light field image as the light field image to be detected.
According to a preferred embodiment of the present invention, the processor 13 further executes a plurality of instructions including:
compressing the light field image.
According to a preferred embodiment of the present invention, the processor 13 further executes a plurality of instructions including:
compressing the light field image by using a vector quantization algorithm to obtain a quantized compressed light field image;
and processing the quantized compressed light field image by utilizing an entropy coding algorithm to obtain a compressed light field image.
According to a preferred embodiment of the present invention, the processor 13 further executes a plurality of instructions including:
acquiring a first training sample by utilizing a web crawler technology, wherein the first training sample comprises positive sample data representing a face picture and negative sample data representing a non-face picture;
and training the face detector by adopting a neural network algorithm according to the first training sample.
According to a preferred embodiment of the present invention, the processor 13 further executes a plurality of instructions including:
acquiring an acquired light field image;
detecting the collected light field image by using the trained face detector to obtain a light field image with a face;
determining the light field image with the face as a second training sample, wherein the second training sample comprises positive sample data representing a clear picture and negative sample data representing a fuzzy picture;
and taking the second training sample as input data of the face detector, and training the ambiguity judgment model by combining a neural network algorithm.
Specifically, the processor 13 may refer to the description of the relevant steps in the embodiment corresponding to fig. 1 for a specific implementation method of the instruction, which is not described herein again.
In the embodiments provided in the present invention, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof.
The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.
Furthermore, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of units or means recited in the system claims may also be implemented by one unit or means in software or hardware. The terms second, etc. are used to denote names, but not any particular order.
Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims (8)
1. A light field image filtering method, the method comprising:
acquiring a light field image to be detected;
detecting the light field image to be detected by using a face detector to obtain a face image;
determining the ambiguity of the face picture by using a trained ambiguity judgment model;
filtering the face picture according to the fuzziness of the face picture;
training the ambiguity resolution model comprises: acquiring an acquired light field image; detecting the collected light field image by using the face detector to obtain a light field image with a face; determining the light field image with the face as a first training sample, wherein the first training sample comprises positive sample data representing a clear picture and negative sample data representing a fuzzy picture; inputting the first training sample into the face detector, randomly initializing parameters corresponding to a full connection layer of a network of the face detector to any value in a range of 0-1, and measuring loss of the full connection layer by using a loss function to obtain the ambiguity judgment model;
training the face detector comprises: acquiring a second training sample by utilizing a web crawler technology, wherein the second training sample comprises positive sample data representing a face picture and negative sample data representing a non-face picture; training the face detector by adopting a neural network algorithm according to the second training sample;
wherein, the acquiring of the light field image to be detected comprises:
acquiring a light field image, performing digital refocusing of a specified depth of field on the light field image to obtain a digital refocused light field image, and determining the digital refocused light field image as the light field image to be detected;
wherein the light field pattern is adjustedDigital refocusing an image to a specified depth of field includes: note LFF (x, y, u, v) is a mathematical description of the light-field image,
wherein L isFFor a given amount of radiation of a ray, u, v are focal planes, x, y are image planes, F denotes the distance between the focal plane and the image plane,
wherein the light field image is represented by the following formula;
if the F is adjusted to be F', the new light field image is expressed as the following formula;
wherein α represents a refocusing factor, BαRepresenting fourier coefficients;
wherein the method further comprises: the electronic equipment normalizes input face picture data to realize unification of data types and formats, the input face picture data is normalized to be 48 multiplied by 48, the electronic equipment respectively inputs the input face picture data into 4 convolution layers of 3 multiplied by 3, a maximum value pooling layer of 3 multiplied by 3 and 1 full connection layer, classification of the input face picture data and prediction of face positions in the input face picture data are finished through a loss function and an Euclidean distance algorithm respectively.
2. The light field image filtering method according to claim 1, wherein prior to digitally refocusing the light field image for a specified depth of field, the method further comprises:
compressing the light field image.
3. The light field image filtering method of claim 2, wherein said compressing the light field image comprises:
compressing the light field image by using a vector quantization algorithm to obtain a quantized compressed light field image;
and processing the quantized compressed light field image by utilizing an entropy coding algorithm to obtain a compressed light field image.
4. A method for face analysis, the method comprising:
acquiring a light field image to be detected;
filtering the light field image to be detected by using the light field image filtering method according to any one of claims 1 to 3 to obtain a reserved human face image;
analyzing and processing the reserved human face picture corresponding to the appointed scene to obtain an analysis result;
and executing the operation corresponding to the specified scene according to the analysis result.
5. The method of analyzing a face according to claim 4, wherein the analyzing the retained face picture corresponding to a designated scene to obtain an analysis result, and the performing an operation corresponding to the designated scene according to the analysis result includes:
identifying an image with a target person from the reserved face picture;
and sending the identified image with the target person to at least one terminal device.
6. The method of analyzing a face according to claim 4, wherein the analyzing the retained face picture corresponding to a designated scene to obtain an analysis result, and performing the operation corresponding to the designated scene according to the analysis result further comprises:
when the face picture is shot by a camera device of a specified vehicle, judging whether a pedestrian exists in the reserved face picture;
and when a pedestrian exists in the face picture, controlling the designated vehicle to brake.
7. An electronic device, characterized in that the electronic device comprises:
a memory storing at least one instruction; and
a processor executing instructions stored in the memory to implement the light field image filtering method of any of claims 1 to 3.
8. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, is capable of implementing the method of any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711423426.8A CN107958231B (en) | 2017-12-25 | 2017-12-25 | Light field image filtering method, face analysis method and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711423426.8A CN107958231B (en) | 2017-12-25 | 2017-12-25 | Light field image filtering method, face analysis method and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107958231A CN107958231A (en) | 2018-04-24 |
| CN107958231B true CN107958231B (en) | 2022-01-11 |
Family
ID=61956924
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711423426.8A Active CN107958231B (en) | 2017-12-25 | 2017-12-25 | Light field image filtering method, face analysis method and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107958231B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108898579B (en) * | 2018-05-30 | 2020-12-01 | 腾讯科技（深圳）有限公司 | Image definition recognition method and device and storage medium |
| CN110866424A (en) * | 2018-08-28 | 2020-03-06 | 天津理工大学 | False face detection method based on 4D light field imaging |
| CN112950525B (en) * | 2019-11-22 | 2024-09-17 | 华为技术有限公司 | Image detection method and device and electronic equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060228039A1 (en) * | 2003-02-28 | 2006-10-12 | Simon Richard A | Method and system for enhancing portrait images that are processed in a batch mode |
| CN103605971A (en) * | 2013-12-04 | 2014-02-26 | 深圳市捷顺科技实业股份有限公司 | Method and device for capturing face images |
| CN105205458A (en) * | 2015-09-16 | 2015-12-30 | 北京邮电大学 | Human face living detection method, device and system |
| CN105631439A (en) * | 2016-02-18 | 2016-06-01 | 北京旷视科技有限公司 | Human face image collection method and device |
-
2017
- 2017-12-25 CN CN201711423426.8A patent/CN107958231B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060228039A1 (en) * | 2003-02-28 | 2006-10-12 | Simon Richard A | Method and system for enhancing portrait images that are processed in a batch mode |
| CN103605971A (en) * | 2013-12-04 | 2014-02-26 | 深圳市捷顺科技实业股份有限公司 | Method and device for capturing face images |
| CN105205458A (en) * | 2015-09-16 | 2015-12-30 | 北京邮电大学 | Human face living detection method, device and system |
| CN105631439A (en) * | 2016-02-18 | 2016-06-01 | 北京旷视科技有限公司 | Human face image collection method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107958231A (en) | 2018-04-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022042049A1 (en) | Image fusion method, and training method and apparatus for image fusion model | |
| KR102172234B1 (en) | Image processing method and apparatus, and electronic device | |
| CN111160202B (en) | Identity verification method, device, equipment and storage medium based on AR equipment | |
| CN111079764B (en) | Low-illumination license plate image recognition method and device based on deep learning | |
| WO2018113523A1 (en) | Image processing method and device, and storage medium | |
| WO2022067656A1 (en) | Image processing method and apparatus | |
| CN110136144B (en) | An image segmentation method, device and terminal device | |
| CN109711364A (en) | A kind of facial image super-resolution reconstruction method, device and computer equipment | |
| CN107958231B (en) | Light field image filtering method, face analysis method and electronic equipment | |
| CN113299073B (en) | Method, device, equipment and storage medium for identifying illegal parking of vehicle | |
| US20210201501A1 (en) | Motion-based object detection method, object detection apparatus and electronic device | |
| CN114140346A (en) | Image processing method and device | |
| CN117132515B (en) | Image processing method and electronic device | |
| CN108040244A (en) | Grasp shoot method and device, storage medium based on light field video flowing | |
| CN117079261A (en) | Campus vehicle cross-domain tracking method, system, equipment and storage medium | |
| JP2022545962A (en) | Fog Recognition Method and Apparatus, Electronic Device, Storage Medium and Computer Program Product | |
| CN113379858A (en) | Image compression method and device based on deep learning | |
| Que et al. | Lightweight and dynamic deblurring for IoT-enabled smart cameras | |
| CN111062337B (en) | People stream direction detection method and device, storage medium and electronic equipment | |
| CN116052047B (en) | Moving object detection method and related equipment | |
| CN114926348B (en) | Device and method for removing low-illumination video noise | |
| CN104125430A (en) | Method and device for detecting video moving objects as well as video monitoring system | |
| CN117876786A (en) | A multi-scale target detection method based on adaptive polarization coded images | |
| CN113628148B (en) | Method and device for reducing noise of infrared image | |
| CN116664694A (en) | Training method of image brightness acquisition model, image acquisition method and mobile terminal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |