Target object identification method and device, equipment, medium and product

Abstract

The invention discloses a target object identification method and device, equipment, a medium and a product. The method comprises the following steps: acquiring a to-be-recognized image; the to-be-recognized image is input into a target neural network model to obtain a face image of a target object, the target neural network model is obtained by iteratively training a convolutional neural network model through a target image sample set, and the target image sample set comprises a to-be-recognized image sample and a standard image sample corresponding to the to-be-recognized image sample; and comparing the face image of the target object with a target image in a database to obtain an identification result. Through the technical scheme of the invention, the influence of environment illumination, face posture change, scene interference items and the like in a face recognition process can be reduced, and the recognition accuracy is improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of face recognition technology, and in particular to a target object recognition method, device, equipment, medium and product. Background Art

[0002] Face recognition technology refers to the use of computer vision algorithms to automatically process acquired facial image sequences to identify whether the image matches the identity information. At present, bank customers are required to conduct real-time verification more and more frequently when handling financial services. Among them, face recognition of customers is a reliable method to verify the customer's identity information and the environment they are in. Therefore, developing a face recognition system with strong recognition capabilities, high real-time performance, and customer-friendly operation is an important measure for banks to optimize branch management, enhance remote service capabilities, and improve customer satisfaction with services.

[0003] However, in the face recognition process based on financial services, there is uncertainty in the environment and operation of the person being recognized. The portrait data obtained by the system is often affected by factors such as ambient lighting, facial posture changes, and scene interference. Simple image feature extraction methods have poor robustness and low recognition accuracy. In view of this, face recognition technology applied to banking service scenarios should take into account the diversity of bank branch scales, the complexity of customer environments, and the unpredictability of customer operations. Therefore, an end-to-end recognition system is needed that takes customer portraits as a salient target and is insensitive to global image information such as ambient light sources, scene miscellaneous items, and posture changes, in order to improve the accuracy and efficiency of face recognition services. Summary of the invention

[0004] Embodiments of the present invention provide a target object recognition method, device, equipment, medium and product to reduce the effects of ambient lighting, facial posture changes, scene interference items, etc. during face recognition and improve recognition accuracy.

[0005] According to one aspect of the present invention, a method for identifying a target object is provided, comprising:

[0006] Obtain an image to be recognized;

[0007] Inputting the image to be identified into a target neural network model to obtain a facial image of a target object, wherein the target neural network model is obtained by iteratively training a convolutional neural network model with a target image sample set, wherein the target image sample set includes: an image sample to be identified and a standard image sample corresponding to the image sample to be identified;

[0008] The facial image of the target object is compared with the target image in the database to obtain a recognition result.

[0009] According to another aspect of the present invention, a target object recognition device is provided, the device comprising:

[0010] An acquisition module, used for acquiring an image to be recognized;

[0011] An input module is used to input the image to be identified into a target neural network model to obtain a facial image of a target object, wherein the target neural network model is obtained by iteratively training a convolutional neural network model with a target image sample set, wherein the target image sample set includes: an image sample to be identified and a standard image sample corresponding to the image sample to be identified;

[0012] The comparison module is used to compare the facial image of the target object with the target image in the database to obtain a recognition result.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the target object recognition method described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the target object recognition method described in any embodiment of the present invention when executed.

[0018] According to another aspect of the present invention, an embodiment of the present invention further provides a computer program product, the computer program product includes a computer program, and when the computer program is executed by a processor, the target object recognition method described in any embodiment of the present invention is implemented.

[0019] The embodiment of the present invention obtains the facial image of the target object by first acquiring the image to be recognized, and then inputting the image to be recognized into the target neural network model, wherein the target neural network model is obtained by iteratively training the convolutional neural network model through the target image sample set, and specifically, the target image sample set includes: the image sample to be recognized and the standard image sample corresponding to the image sample to be recognized, and finally the facial image of the target object is compared with the target image in the database to obtain the recognition result. Through the technical solution of the present invention, the influence of ambient lighting, facial posture changes, scene interference items, etc. in the face recognition process can be reduced, and the recognition accuracy rate can be improved.

[0020] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 is a flow chart of a target object recognition method in an embodiment of the present invention;

[0023] Figure 2 is a schematic structural diagram of a target object recognition device in an embodiment of the present invention;

[0024] Figure 3 It is a structural schematic diagram of an electronic device for implementing the target object recognition method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and their meanings are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0028] Embodiment 1

[0029] Figure 1 is a flow chart of a target object recognition method in an embodiment of the present invention. This embodiment is applicable to the case of target object recognition. The method can be executed by a target object recognition device in an embodiment of the present invention. The device can be implemented in software and / or hardware. Figure 1 As shown, the method specifically comprises the following steps:

[0030] S101: Obtain an image to be recognized.

[0031] It should be noted that the image to be recognized may be an image to be subjected to face recognition.

[0032] Exemplarily, in this embodiment, the image to be identified may be an image containing a face, taken when a bank customer needs to perform real-time identity verification when handling financial services.

[0033] S102: Input the image to be recognized into the target neural network model to obtain the facial image of the target object.

[0034] In this embodiment, the target neural network model may be a neural network model for extracting a face image from an image to be recognized. Exemplarily, the target neural network model may be, for example, a trained convolutional neural network model.

[0035] The target neural network model is obtained by iteratively training the convolutional neural network model with a target image sample set, and the target image sample set includes: image samples to be identified and standard image samples corresponding to the image samples to be identified.

[0036] It should be noted that the target image sample set may be a set of image samples used to train a convolutional neural network model. The image samples to be identified may be pre-collected images containing human faces. Specifically, the image samples to be identified may be sampled from a face image video, for example. The standard image samples may be standard face images corresponding to the image samples to be identified, which may be used for identity comparison.

[0037] The target object may be a user to be authenticated.

[0038] In this embodiment, the image to be identified may be preprocessed to obtain an image to be identified after feature enhancement, and then the image to be identified is input into a trained target neural network model to obtain a facial image of the target object.

[0039] Furthermore, the image to be identified is preprocessed to obtain the image to be identified after feature enhancement, including: performing a geometric radiation transformation on the image to be identified; based on the geometric radiation transformation, increasing the pixel change value of the dark area of ​​the image to be identified; based on the pixel change value, calculating the new image pixel value of the image to be identified, highlighting the structure of the dark area, and completing image enhancement.

[0040] In this embodiment, the image to be identified is preprocessed by geometric radiation transformation to obtain the image to be identified after feature enhancement, including: feature enhancement under contrast limitation, thereby relaxing the brightness constancy assumption and improving the problem that too dark areas cannot be perceived in endoscopic scenes. After image enhancement, except for isolated discontinuous boundaries caused by lighting conditions and surface reflectivity mutations, the pixel change value of the entire scene is expected to be smooth and can reflect more scene structure information. An expanded 3×3 convolution with a sampling interval is set inside the kernel, which can increase the receptive field inside the convolution kernel, so that the convolution layer has a larger receptive field and achieves the purpose of global information perception. After obtaining the feature maps under the channel and spatial receptive fields, they need to be spliced ​​and fused using 1×1 convolution, assisted by residual connections to prevent gradient disappearance.

[0041] S103: Compare the facial image of the target object with the target image in the database to obtain a recognition result.

[0042] It should be noted that the target image may be a standard face image stored in a database. The recognition result may be to identify a standard image corresponding to the target object from all standard images stored in the database based on the facial image of the target object. The identity information of the target object may be subsequently verified based on the standard image corresponding to the target object.

[0043] Specifically, the facial image of the target object output by the target neural network model is compared with the standard facial image stored in the database, and the comparison result is output.

[0044] The embodiment of the present invention obtains the facial image of the target object by first acquiring the image to be recognized, and then inputting the image to be recognized into the target neural network model, wherein the target neural network model is obtained by iteratively training the convolutional neural network model through the target image sample set, and specifically, the target image sample set includes: the image sample to be recognized and the standard image sample corresponding to the image sample to be recognized, and finally the facial image of the target object is compared with the target image in the database to obtain the recognition result. Through the technical solution of the present invention, the influence of ambient lighting, facial posture changes, scene interference items, etc. in the face recognition process can be reduced, and the recognition accuracy rate can be improved.

[0045] Optionally, iteratively training the convolutional neural network model using the target image sample set includes:

[0046] Build a convolutional neural network model.

[0047] The image samples to be identified in the target sample set are input into the convolutional neural network model to obtain the predicted image.

[0048] The predicted image may be a face image output by a convolutional neural network model, which is obtained by processing the image samples to be identified in the target sample set.

[0049] In this embodiment, the convolutional neural network model can be divided into two parts: an encoder and a decoder. In actual operation, the image sample to be identified in the target sample set is input into the encoder, and then the output of the convolutional neural network encoder is restored to resolution through the decoder to obtain a face image that integrates global information, that is, a predicted image.

[0050] Update the parameters of the convolutional neural network model according to the predicted image and the standard image samples corresponding to the image samples to be identified.

[0051] Specifically, the parameters of the convolutional neural network encoder and decoder can be updated according to the sample training results.

[0052] Return to execute the operation of inputting the image samples to be identified in the target sample set into the convolutional neural network model to obtain the predicted image until the target neural network model is obtained.

[0053] Optionally, updating the parameters of the convolutional neural network model according to the predicted image and the standard image samples corresponding to the image samples to be identified includes:

[0054] A first loss value is determined according to the predicted image and the standard image samples corresponding to the image samples to be identified.

[0055] In this embodiment, the first loss value may be an image similarity structure loss value obtained by performing image structure loss calculation on the predicted image and the standard image sample corresponding to the image sample to be identified.

[0056] It is known that the Structural Similarity Index Measure (SSIM) is an indicator for measuring the visual similarity between two images. SSIM is calculated based on the structural information, brightness and contrast of the image, and it can better reflect the human eye's perception of changes in image quality.

[0057] Specifically, the image similarity structural loss is a combination of the absolute error of image pixels and the SSIM calculation formula. SSIM calculation is an indicator used to measure the structural similarity of images, and its value is in the range of [0, 1]. The closer to 1, the more similar the two images are. Absolute error is another commonly used image similarity indicator, which represents the difference between two images. This embodiment uses the two to jointly measure the difference between the predicted image and the standard image sample corresponding to the image sample to be identified.

[0058] Obtain the depth map samples corresponding to the image samples to be identified in the target sample set.

[0059] The depth map sample may be a depth map corresponding to the image sample to be identified, and the depth map may record the distance information from each pixel in the scene to the camera or sensor.

[0060] Specifically, a depth sensor or a depth estimation neural network may be used to obtain depth map samples corresponding to image samples to be identified in a target sample set.

[0061] A second loss value is determined based on the depth map samples.

[0062] In this embodiment, the second loss value may be a value obtained by performing edge smoothing loss calculation on the depth map sample.

[0063] It can be known that the edge smoothing loss is an exponential function calculation formula, which takes the depth map sample gradient as the exponent of the exponential function with the natural logarithm as the base. According to the properties of the exponential function, when the depth map sample pixels change greatly, the exponential product term of the loss will become smaller, reducing the impact on the generation of the global information fusion portrait image, and reducing the edge sharp points in the process of the network predicting the portrait image from the original image. Among them, edge sharpness refers to the pixel points where the image pixel value changes too much in certain areas of the depth map sample.

[0064] Update the parameters of the convolutional neural network model according to the first loss value and the second loss value.

[0065] Specifically, the parameters of the convolutional neural network encoder and decoder are updated according to the image similarity structure loss and edge smoothing loss. The convolutional neural network parameters after the cycle are used to extract the face image that integrates the global information, and the image is compared with the standard face image stored in the database, and the comparison result is output.

[0066] In actual operation, based on the acquired face image video, a depth sensor or a depth estimation neural network can be used to obtain a depth image corresponding to each frame of the video. Based on a preset sampling frame interval, a certain number of images are selected from the total sequence of video images and the total sequence of depth estimation images as image samples to be identified and depth map samples, respectively.

[0067] In this embodiment, the depth estimation neural network may be a convolutional neural network, which may include an encoder and a decoder, wherein the encoder may be, for example, a feature extraction model, and the decoder may be, for example, a resolution recovery model. Further, the specific implementation process of using the depth estimation neural network to obtain the depth map sample corresponding to the image sample to be identified in the target sample set is as follows:

[0068] Optionally, obtaining a depth map sample corresponding to the image sample to be identified in the target sample set includes:

[0069] The image samples to be identified in the target sample set are input into the feature extraction model to obtain the image feature tensor corresponding to the image samples to be identified.

[0070] In this embodiment, the feature extraction model may be a convolutional neural network encoder for extracting image features, wherein the image feature tensor may be the image features corresponding to the image samples to be identified in the target sample set output by the feature extraction model.

[0071] Specifically, the image samples to be identified in the target sample set are input into a pre-trained feature extraction model to obtain an image feature tensor corresponding to the image samples to be identified.

[0072] The image feature tensor is input into the resolution recovery model to obtain the depth image sample corresponding to the image sample to be identified.

[0073] In this embodiment, the resolution recovery model can be a convolutional neural network decoder, which is used to restore image features into a depth estimation image.

[0074] Specifically, the image feature tensor is input into a pre-trained resolution restoration model to obtain a depth estimation image corresponding to the corresponding face image and image feature tensor. The depth estimation image represents the distance relationship between the face and other instances in the image and the camera, and is used to highlight the saliency of the face in the entire image.

[0075] Optionally, the convolutional neural network model includes: a global feature information perception sub-model and a scale fusion recovery sub-model. The global feature information perception sub-model includes: a feature extraction network module and a spatial information fusion module.

[0076] The image samples to be identified in the target sample set are input into the convolutional neural network model to obtain the predicted image, including:

[0077] The image samples to be identified in the target sample set are input into the feature extraction network module to obtain a first image feature vector set.

[0078] The first image feature vector set may be a set of image feature vectors corresponding to the image samples to be identified in the target sample set output by the feature extraction network module.

[0079] Specifically, the feature extraction network module is a convolutional neural network encoder, which is used to obtain a first image feature vector set.

[0080] The first image feature vector set is input into the spatial information fusion module to obtain a second image feature vector set.

[0081] The second image feature vector set is obtained by performing spatial information fusion processing on the first image feature vector set.

[0082] Specifically, the first image feature vector set is input into the spatial information fusion module, and the spatial information fusion processing is performed on the first image feature vector set to obtain the second image feature vector set.

[0083] Furthermore, the spatial information fusion module performs spatial information fusion processing on the first image feature vector set to obtain the second image feature vector set, which can be expressed as:

[0084]

[0085] Among them, X S Represents the second image feature vector set of the final output, Represents a convolution operation with a kernel spacing of R, a length and a width of 3, C 1×1 represents a convolution operation with a length and width of 1, X i represents the image feature vector set output in step i, X in Represents the first image feature vector set.

[0086] The second image feature vector set is input into the scale fusion recovery sub-model to obtain a third image feature vector set.

[0087] The third image feature vector set may be a set of image feature vectors output by the scale fusion restoration sub-model. In this embodiment, the scale fusion restoration sub-model is essentially a multi-scale fusion convolutional neural network decoder.

[0088] Specifically, the second image feature vector set is input into a multi-scale fusion convolutional neural network decoder, and based on the multi-scale fusion convolutional neural network decoder, a multi-scale feature face image is generated. Based on the multi-scale feature face image, a residual module is used to fuse feature images of different scales to obtain a third image feature vector set.

[0089] A deconvolution operation is performed on the third image feature vector set to obtain a predicted image.

[0090] Specifically, based on the third image feature vector set, pooling and deconvolution operations are performed on it to restore it to a maximum size to obtain a fourth image feature vector set, that is, to obtain a predicted image.

[0091] In this embodiment, a pyramid fusion learning method is used, the purpose of which is to reduce the loss of the resolution recovery process in the decoder module. In each frame, the decoder generates a progressive multi-scale feature map during the upsampling process. The features extracted at different stages are upsampled and fused by the scale pyramid method, and finally a feature map of the input size is obtained, so as to more accurately predict the pixel value of the image sample to be identified. The multi-scale fusion mechanism similar to the RefineNet network uses residual connections to form a long link with the four modules of the ResNet encoder to ensure that the gradient can be passed to the entire network. The multi-scale fusion head of the pyramid fusion learning uses residual connections to establish a long-distance information flow. In the feature map of each scale, the network uses two residual blocks to extract features. Then, the deconvolution operation is used to upsample the feature map of each scale to the same size. In this step, the feature map obtained by the encoder at each step uses multiple residual blocks for fusion, and these residual blocks establish long-range connections between multiple levels, so as to better capture information at different levels. The 2-5 layer resolution feature maps extracted by the ResNet-18 residual network are restored to the maximum size through global information perception and multi-scale context aggregation, and finally passed to subsequent processing through pooling, convolution and weighted operation output.

[0092] Optionally, the facial image of the target object is compared with the target image in the database to obtain a recognition result, including:

[0093] The image structure similarity between the target object's facial image and the target image in the database is calculated to obtain a similarity score.

[0094] Specifically, the facial image of the target object is subjected to image structure similarity calculation with the standard facial image in the database to obtain a similarity score.

[0095] The recognition result is determined based on the similarity score.

[0096] In actual operation, a facial image whose similarity score is greater than a preset threshold (wherein the preset threshold may be pre-set by the user based on actual conditions or experience, and this embodiment does not impose a specific numerical limit on this) and whose similarity score is closest to the preset threshold may be extracted, or a facial image with the largest similarity score may be extracted from a set of similarity scores as the recognition result corresponding to the image to be recognized, thereby verifying the user's identity information.

[0097] The technical solution of the embodiment of the present invention uses geometric radiation transformation during image preprocessing to increase the pixel change value of the dark area of ​​the image to be identified, highlight the structure of the dark area, and complete image enhancement; during face recognition, in addition to setting the camera to collect images containing faces, a depth map of the current scene is obtained through a depth sensor or a depth estimation convolutional neural network, and the depth map will be used as an auxiliary training image to complete the face recognition process; in the face recognition process, a global information perception and context feature fusion method is proposed to further improve the effect of feature aggregation and perceive the global information of the image other than the portrait during the face recognition process. The technical solution of the embodiment of the present invention can perceive and identify the global information of the image to reduce the impact of changes in light intensity, changes in facial posture, and other scene object instances in the global information, and improve the recognition accuracy.

[0098] Embodiment 2

[0099] Figure 2 is a schematic diagram of the structure of a target object recognition device in an embodiment of the present invention. This embodiment is applicable to the case of target object recognition. The device can be implemented in software and / or hardware. The device can be integrated in any device that provides the function of target object recognition, such as Figure 2 As shown, the target object recognition device specifically includes: an acquisition module 201, an input module 202 and a comparison module 203.

[0100] Wherein, the acquisition module 201 is used to acquire the image to be identified;

[0101] An input module 202 is used to input the image to be identified into a target neural network model to obtain a facial image of a target object, wherein the target neural network model is obtained by iteratively training a convolutional neural network model with a target image sample set, wherein the target image sample set includes: an image sample to be identified and a standard image sample corresponding to the image sample to be identified;

[0102] The comparison module 203 is used to compare the facial image of the target object with the target image in the database to obtain a recognition result.

[0103] Optionally, the input module 202 includes:

[0104] Establish a unit for establishing a convolutional neural network model;

[0105] An input unit, used for inputting the image samples to be identified in the target sample set into the convolutional neural network model to obtain a predicted image;

[0106] An updating unit, used for updating the parameters of the convolutional neural network model according to the predicted image and the standard image sample corresponding to the image sample to be identified;

[0107] The execution unit is used to return to the operation of inputting the image samples to be identified in the target sample set into the convolutional neural network model to obtain a predicted image until the target neural network model is obtained.

[0108] Optionally, the updating unit includes:

[0109] A first determining subunit, configured to determine a first loss value according to the predicted image and a standard image sample corresponding to the image sample to be identified;

[0110] An acquisition subunit, used to acquire a depth map sample corresponding to the image sample to be identified in the target sample set;

[0111] A second determining subunit, configured to determine a second loss value according to the depth map sample;

[0112] An updating subunit, used to update the parameters of the convolutional neural network model according to the first loss value and the second loss value.

[0113] Optionally, the acquisition subunit is specifically used for:

[0114] Inputting the image samples to be identified in the target sample set into a feature extraction model to obtain image feature tensors corresponding to the image samples to be identified;

[0115] The image feature tensor is input into a resolution recovery model to obtain a depth image sample corresponding to the image sample to be identified.

[0116] Optionally, the convolutional neural network model includes: a global feature information perception sub-model and a scale fusion recovery sub-model; the global feature information perception sub-model includes: a feature extraction network module and a spatial information fusion module;

[0117] The input unit is specifically used for:

[0118] Inputting the image samples to be identified in the target sample set into the feature extraction network module to obtain a first image feature vector set;

[0119] Inputting the first image feature vector set into the spatial information fusion module to obtain a second image feature vector set, where the second image feature vector set is obtained after performing spatial information fusion processing on the first image feature vector set;

[0120] Inputting the second image feature vector set into the scale fusion restoration sub-model to obtain a third image feature vector set;

[0121] A deconvolution operation is performed on the third image feature vector set to obtain a predicted image.

[0122] Optionally, the comparison module 203 is specifically used for:

[0123] Calculating image structure similarity between the facial image of the target object and the target image in the database to obtain a similarity score;

[0124] A recognition result is determined according to the similarity score.

[0125] The above-mentioned product can execute the target object recognition method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0126] Embodiment 3

[0127] Figure 3 A schematic diagram of the structure of an electronic device 30 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0128] like Figure 3As shown, the electronic device 30 includes at least one processor 31, and a memory connected to the at least one processor 31, such as a read-only memory (ROM) 32, a random access memory (RAM) 33, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 31 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 32 or the computer program loaded from the storage unit 38 to the random access memory (RAM) 33. In the RAM 33, various programs and data required for the operation of the electronic device 30 can also be stored. The processor 31, the ROM 32, and the RAM 33 are connected to each other through a bus 34. An input / output (I / O) interface 35 is also connected to the bus 34.

[0129] A number of components in the electronic device 30 are connected to the I / O interface 35, including: an input unit 36, such as a keyboard, a mouse, etc.; an output unit 37, such as various types of displays, speakers, etc.; a storage unit 38, such as a disk, an optical disk, etc.; and a communication unit 39, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 39 allows the electronic device 30 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0130] The processor 31 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 31 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 31 performs the various methods and processes described above, such as the target object recognition method:

[0131] Obtain an image to be recognized;

[0132] Inputting the image to be identified into a target neural network model to obtain a facial image of a target object, wherein the target neural network model is obtained by iteratively training a convolutional neural network model with a target image sample set, wherein the target image sample set includes: an image sample to be identified and a standard image sample corresponding to the image sample to be identified;

[0133] The facial image of the target object is compared with the target image in the database to obtain a recognition result.

[0134] In some embodiments, the target object recognition method can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 38. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 30 via the ROM 32 and / or the communication unit 39. When the computer program is loaded into the RAM 33 and executed by the processor 31, one or more steps of the target object recognition method described above can be performed. Alternatively, in other embodiments, the processor 31 can be configured to perform the target object recognition method in any other appropriate manner (for example, by means of firmware).

[0135] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0136] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0137] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0138] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0139] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0140] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0141] In one embodiment, the embodiment of the present invention further includes a computer program product, the computer program product includes a computer program, and when the computer program is executed by a processor, the target object recognition method of any embodiment of the present invention is implemented.

[0142] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0143] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0144] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A target object recognition method, characterized in that: include: Obtain an image to be recognized; Inputting the image to be identified into a target neural network model to obtain a facial image of a target object, wherein the target neural network model is obtained by iteratively training a convolutional neural network model with a target image sample set, wherein the target image sample set includes: an image sample to be identified and a standard image sample corresponding to the image sample to be identified; The facial image of the target object is compared with the target image in the database to obtain a recognition result.

2. The method according to claim 1, characterized in that Iteratively training the convolutional neural network model through the target image sample set includes: Build a convolutional neural network model; Inputting the image samples to be identified in the target sample set into the convolutional neural network model to obtain a predicted image; Updating the parameters of the convolutional neural network model according to the predicted image and the standard image sample corresponding to the image sample to be identified; Return to executing the operation of inputting the image samples to be identified in the target sample set into the convolutional neural network model to obtain a predicted image until the target neural network model is obtained.

3. The method according to claim 2, characterized in that Updating the parameters of the convolutional neural network model according to the predicted image and the standard image sample corresponding to the image sample to be identified, including: Determine a first loss value according to the predicted image and a standard image sample corresponding to the image sample to be identified; Obtaining depth map samples corresponding to the image samples to be identified in the target sample set; determining a second loss value based on the depth map samples; Update the parameters of the convolutional neural network model according to the first loss value and the second loss value.

4. The method according to claim 3, characterized in that Obtaining a depth map sample corresponding to the image sample to be identified in the target sample set, including: Inputting the image samples to be identified in the target sample set into a feature extraction model to obtain image feature tensors corresponding to the image samples to be identified; The image feature tensor is input into a resolution recovery model to obtain a depth image sample corresponding to the image sample to be identified.

5. The method according to claim 2, characterized in that: The convolutional neural network model includes: a global feature information perception sub-model and a scale fusion recovery sub-model; the global feature information perception sub-model includes: a feature extraction network module and a spatial information fusion module; Inputting the image samples to be identified in the target sample set into the convolutional neural network model to obtain a predicted image, including: Inputting the image samples to be identified in the target sample set into the feature extraction network module to obtain a first image feature vector set; Inputting the first image feature vector set into the spatial information fusion module to obtain a second image feature vector set, where the second image feature vector set is obtained after performing spatial information fusion processing on the first image feature vector set; Inputting the second image feature vector set into the scale fusion restoration sub-model to obtain a third image feature vector set; A deconvolution operation is performed on the third image feature vector set to obtain a predicted image.

6. The method according to claim 1, characterized in that Comparing the facial image of the target object with the target image in the database to obtain a recognition result includes: Calculating image structure similarity between the target object's facial image and a target image in a database to obtain a similarity score; A recognition result is determined according to the similarity score.

7. A target object recognition device, characterized in that: include: An acquisition module, used for acquiring an image to be recognized; An input module is used to input the image to be identified into a target neural network model to obtain a facial image of a target object, wherein the target neural network model is obtained by iteratively training a convolutional neural network model with a target image sample set, wherein the target image sample set includes: an image sample to be identified and a standard image sample corresponding to the image sample to be identified; The comparison module is used to compare the facial image of the target object with the target image in the database to obtain a recognition result.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the target object recognition method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the target object recognition method according to any one of claims 1 to 6 when executed.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the target object recognition method according to any one of claims 1 to 6.

Images