WO2024255425A1 - Image acquisition - Google Patents

Abstract

One or more embodiments of the present description disclose an image acquisition method and apparatus, and an electronic device. The method comprises: acquiring an image captured for a target object; then when the image does not satisfy a preset image quality condition, inputting the image into a pre-trained reason generation model, determining a location, which does not satisfy the preset image quality condition, in the image by means of the reason generation model, and generating image recapture guide information for the determined location, wherein the reason generation model is a model used for detecting a reason causing the image not to satisfy the preset image quality condition, and providing a corresponding improvement suggestion for a detection result; and finally performing image capture on the target object again on the basis of the image recapture guide information.

Description

Image acquisition Technical Field

This document relates to the field of image recognition technology, and in particular to an image acquisition method, device and electronic equipment.

Background Art

In the field of image recognition, when capturing an image of a target object, it often happens that the image obtained in one capture fails to meet user satisfaction. Take the eKYC (electronic Know Your Customer, enterprise electronic identity authentication information system) process as an example. In the eKYC process, users need to take photos of their ID and face respectively to complete identity authentication. If the quality of the pictures taken by users is low (such as blur, highlights, etc.), it will affect the accuracy of ID and face recognition, resulting in the user being unable to successfully open an account. Therefore, how to capture images to obtain high-quality images that meet user satisfaction is a key issue in improving user experience.

At present, when the image obtained by one shooting cannot meet the user's satisfaction, the method usually adopted is to continue shooting until the image obtained meets the user's needs. Therefore, it is necessary to provide a more efficient image acquisition method that can reduce the number of repeated shootings.

Summary of the invention

On the one hand, one or more embodiments of the present specification provide an image acquisition method, comprising: acquiring an image taken of a target object; when the image does not meet a preset image quality condition, inputting the image into a pre-trained cause generation model, determining the position of the image that does not meet the preset image quality condition through the cause generation model, and generating image reshooting guidance information for the determined position, wherein the cause generation model is a model for detecting the cause of the image not meeting the preset image quality condition and providing corresponding improvement suggestions for the detection result; and re-taking the image of the target object according to the image reshooting guidance information.

On the other hand, one or more embodiments of the present specification provide an image acquisition device, including: an image acquisition module, which acquires an image shot for a target object; an image reshooting guidance information generation module, which, when the image does not meet a preset image quality condition, inputs the image into a pre-trained cause generation model, determines the position of the image that does not meet the preset image quality condition through the cause generation model, and generates image reshooting guidance information for the determined position, wherein the cause generation model is a model for detecting the cause causing the image to not meet the preset image quality condition and providing corresponding improvement suggestions for the detection result; and a control module, which re-shoots the image of the target object according to the image reshooting guidance information.

On the other hand, one or more embodiments of the present specification provide an electronic device, comprising: a processor; and a memory arranged to store computer-executable instructions, wherein when the executable instructions are executed, the processor can: obtain an image taken of a target object; when the image does not meet a preset image quality condition, input the image into a pre-trained cause generation model, determine the position of the image that does not meet the preset image quality condition through the cause generation model, and generate image reshooting guidance information for the determined position, wherein the cause generation model is a model for detecting the cause of the image not meeting the preset image quality condition and providing corresponding improvement suggestions for the detection result; and re-take an image of the target object according to the image reshooting guidance information.

On the other hand, one or more embodiments of the present specification provide a storage medium for storing computer-executable instructions, which implement the following process when executed by a processor: obtaining an image taken of a target object; when the image does not meet a preset image quality condition, inputting the image into a pre-trained cause generation model, determining the position of the image that does not meet the preset image quality condition through the cause generation model, and generating image reshooting guidance information for the determined position, wherein the cause generation model is a model for detecting the cause of the image not meeting the preset image quality condition and providing corresponding improvement suggestions for the detection result; and re-taking the image of the target object according to the image reshooting guidance information.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate one or more embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in one or more embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

FIG1 is a schematic flow chart of an image acquisition method according to an embodiment of the present specification;

FIG2 is a schematic flow chart of an image acquisition method according to another embodiment of the present specification;

FIG3 is a schematic flow chart of an image acquisition method according to another embodiment of the present specification;

FIG4 is a schematic diagram of an image acquisition method according to an embodiment of the present specification;

FIG5 is a schematic block diagram of an image acquisition device according to an embodiment of the present specification;

FIG. 6 is a schematic block diagram of an electronic device according to an embodiment of the present disclosure.

DETAILED DESCRIPTION

One or more embodiments of the present specification provide an image acquisition method, device, and electronic device to solve the current problems.

In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the following will be combined with the drawings in one or more embodiments of this specification to clearly and completely describe the technical solutions in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.

As shown in FIG1 , the embodiment of this specification provides an image acquisition method, and the execution subject of the method can be a terminal device, which can be a certain terminal device such as a mobile phone, a tablet computer, or a computer device such as a laptop or a desktop computer, or an IoT device (specifically, a smart watch, a vehicle-mounted device, etc.). The method can specifically include the following steps.

In step S102, an image captured of a target object is acquired.

The method in one or more embodiments of this specification can be applied to scenarios where certain quality requirements are placed on the captured images, such as: capturing scene images with a mobile phone or smart camera, capturing face images with a mobile phone or smart camera, capturing ID images, etc. The target object in an embodiment of this specification can be an ID, a face, or a scene, etc. In actual applications, the target object can also include multiple, for example, the target object includes two, namely an ID and a face, etc.

Taking the eKYC process as an example, the user needs to take photos of the ID and the face respectively to complete the identity verification. According to the above step S102, the target object may include the ID and the face. The ID and the face may be photographed respectively by the camera component of the terminal device to obtain the corresponding ID image and face image. Alternatively, the ID and the face may be captured by the camera component of the terminal device to obtain an image containing both the ID and the face.

In step S104, when the image does not meet the preset image quality conditions, the image is input into a pre-trained cause generation model, the cause generation model is used to determine the position in the image that does not meet the preset image quality conditions, and generate image reshooting guidance information for the determined position. The cause generation model is a model for detecting the reasons why the image does not meet the preset image quality conditions, and providing corresponding improvement suggestions for the detection results.

The preset image quality conditions are determined according to the user's shooting requirements for the target object. For example, when shooting a document or a face in the eKYC process, the preset image quality conditions may include one or more of the following: the shot image reaches a preset pixel size, the shot image reaches a preset resolution, and the shot image reaches a set exposure parameter. Accordingly, images that do not meet the preset image quality conditions usually include: overexposure of highlights (i.e., the highlight area is too bright). The images include images with an exposure higher than a preset exposure threshold value, images that are too dark (i.e., image brightness lower than a preset brightness threshold value and/or image contrast lower than a preset contrast threshold value, etc.), and images that are blurred (i.e., images with pixels lower than a preset pixel threshold value, etc.).

It should be noted that the image obtained through step S102 may or may not meet the preset image quality conditions. There are many ways to determine whether the image meets the preset image quality conditions, such as: pre-setting multiple image quality conditions, and determining whether each image quality condition is met by comparing one by one, so as to determine whether the image meets the preset drawing quality conditions. Alternatively, based on the image quality conditions, a corresponding algorithm (such as a classification algorithm or a random forest algorithm, etc.) can be selected, and a corresponding model can be constructed using the selected algorithm, and the model can be used to determine whether the captured image meets the preset image quality conditions.

In one embodiment of the present specification, a cause generation model is pre-trained. The cause generation model determines the specific location where the quality problem occurs in the image taken last time, detects the cause of the quality problem, and finally generates image reshoot guidance information for the specific location, thereby providing targeted improvement suggestions for the next image shooting. The input data of the cause generation model is an image that does not meet the preset image quality conditions, which may specifically include the image, the information contained in the image, etc. The output result of the cause generation model is the image reshoot guidance information, which can be presented in text form, for example: "The highlight in the upper left corner of the certificate is overexposed, please adjust the highlight area in the upper left corner when shooting", "The overall image is blurred, please keep the phone still when shooting", etc., and the image reshoot guidance information can also be played in the form of voice broadcast.

In one implementation, the image reshooting guidance information may at least include: specific locations of quality issues on corresponding shooting objects and shooting suggestions, and may also include reasons for unqualified quality, etc.

In one embodiment of the present specification, the pre-trained cause generation model can adopt a training model from image to text, and the training model can be constructed based on a neural network, such as by a convolutional neural network, by a transformer, etc. The loss function of the training cause generation model can adopt a standard loss function in text generation.

In one embodiment of the present specification, a method for determining locations in an image that do not meet preset image quality conditions may be based on a generation model that may use a method of overall comparison of the captured image with a standard image, or may use an image segmentation method, i.e., using a segmentation algorithm to determine which pixel locations in the captured image do not meet a preset pixel size.

In step S106, the target object is re-imaged according to the image re-shooting guidance information.

After the image reshooting guidance information is generated in step S104, the generated image reshooting guidance information can be used to guide the user to reshoot the image of the target object in a targeted manner, thereby completing the image acquisition operation with fewer shots, which is beneficial to improving image shooting efficiency.

The embodiment of the present specification provides an image acquisition method, firstly acquiring an image shot for a target object, then determining whether the image meets a preset image quality condition, and when the image does not meet the preset image quality condition, inputting the image into a pre-trained cause generation model, determining the position in the image that does not meet the preset image quality condition through the cause generation model, and generating image reshoot guidance information for the determined position, and finally reshooting the target object according to the generated image reshoot guidance information, thereby completing image acquisition. When the image does not meet the preset image quality condition, the pre-trained cause generation model can detect the cause of the image not meeting the preset image quality condition in the last shooting process, obtain the specific cause of the quality problem, and locate the position in the last shot image that does not meet the preset image quality condition, obtain the specific position that causes the quality problem, and finally generate image reshoot guidance information for the determined position to feedback to the user, thereby providing targeted improvement suggestions for the next shooting, which can effectively reduce the number of repeated shootings, improve image shooting efficiency, and thus improve user experience.

Furthermore, there may be multiple ways to train the cause generation model in the above step S102. An optional processing method is provided below. For details, please refer to the following steps S1022-S1024.

In step S1022, a plurality of image samples taken for different shooting objects, the reason why each image sample does not meet the preset image quality condition, and the position of each image sample that does not meet the preset image quality condition are obtained.

In implementation, the plurality of image samples may include one or more of the following images: historical images of different photographed subjects, original images of each photographed subject that do not meet preset image quality conditions, and images simulated based on the original images of each photographed subject.

In one embodiment, multiple image samples acquired for different shooting objects, the reason why each image sample does not meet the preset image quality conditions, and the position of each image sample that does not meet the preset image quality conditions can be directly input into the cause generation model as input data for model training.

In another implementation, the processing of step S1022 may be performed as the following steps A1 and A2.

Step A1: Acquire multiple image samples taken for different objects.

Step A2: respectively determining the reason why each image sample does not meet the preset image quality condition and the position of each image sample that does not meet the preset image quality condition.

It can be seen from the above steps A1 and A2 that in another embodiment, multiple image samples taken for different shooting objects can be first obtained, and the multiple image samples can be used as input data for model training. Then, the cause generation model determines that each image sample does not meet the preset The reasons for the image quality conditions and the positions of each image sample that do not meet the preset image quality conditions are determined, and then the network layer for generating reason text set in the reason generation model is used to generate corresponding image reshooting guidance information based on the reasons why each image sample does not meet the preset image quality conditions and the positions of each image sample that do not meet the preset image quality conditions.

In another embodiment, the reasons why the image sample does not meet the preset image quality conditions include: a first quality problem and a second quality problem, and the first quality problem is an image quality problem in which the image granularity is greater than or equal to a preset image granularity threshold (or called a large-granularity quality problem), and the second quality problem is an image quality problem in which the image granularity is less than a preset image granularity threshold (or called a fine-granularity quality problem). The processing of obtaining the position in each image sample that does not meet the preset image quality conditions in the above step S1022 can be executed as the following steps B1 and B2.

Step B1: When the reason why the image sample does not meet the preset image quality condition is the first quality problem, the position of each image sample that does not meet the preset image quality condition is determined based on the target detection algorithm.

Step B2: When the reason why the image sample does not meet the preset image quality condition is the second quality problem, the position of each image sample that does not meet the preset image quality condition is identified based on a semantic segmentation algorithm.

Among them, the target detection algorithm is an algorithm for finding the target of interest from the image, and classifying and locating the found target of interest. The target of interest can be roughly framed out by the target detection algorithm. The embodiment of this specification does not limit the specific target detection algorithm. In step B1, the TwoStage detection algorithm or the OneStage detection algorithm based on deep learning can be used. The semantic segmentation algorithm is an algorithm that classifies each pixel in the image, thereby dividing the image into multiple regions containing different categories of information. The semantic segmentation algorithm can accurately depict the outline of the target, and can provide coordinate information and classification information in each pixel. The semantic segmentation algorithm can not only predict the position and category of the target object in the image, but also depict the boundaries between different types of target objects. The embodiment of this specification does not limit the specific semantic segmentation algorithm. In step B2, the FCN (Fully Convolutional Network) algorithm, the CRF (Conditional Random Field algorithm) algorithm, etc. can be used.

It can be seen from the above steps B1 and B2 that when the reason why the image sample does not meet the preset image quality conditions is a large-scale quality problem, the position of the image sample that does not meet the preset image quality conditions can be determined based on the result determined by the target detection algorithm; when the reason why the image sample does not meet the preset image quality conditions is a fine-grained quality problem, the semantic segmentation algorithm can be used to identify the pixel points at which positions in the image sample do not meet the requirements, thereby identifying the position of the image sample that does not meet the preset image quality conditions. Different algorithms can be used to determine the position of each image sample that does not meet the preset image quality conditions according to the different reasons why the image sample does not meet the preset image quality conditions. The location of the quality conditions makes the positioning method for image quality problems more flexible, which can not only effectively save resources, but also efficiently and accurately obtain positioning results that meet the requirements.

In step S1024, based on multiple images, the reasons why each image sample does not meet the preset image quality conditions, and the positions in each image sample that do not meet the preset image quality conditions, the cause generation model is trained using a preset first loss function to obtain a trained cause generation model.

In one embodiment, the preset first loss function adopts a cross entropy loss function. The preset first loss function is used for cause generation model training, which belongs to a supervised training method. It can not only improve the efficiency of model training, but also improve the accuracy of model training output results.

Further, as shown in FIG. 2 , when the image acquisition method in the embodiment of this specification is used in a blockchain system based on electronic identity authentication information, the image acquisition method may include the following steps S202 - S208 .

S202: Collect an image of a designated certificate of the target user and/or an image of the target user taken for the target user to access the blockchain system based on electronic identity authentication information, and use the image of the designated certificate and/or the image of the target user taken as the image taken of the target object.

In implementation, in order to access the blockchain system based on electronic identity authentication information, the target user needs to take a corresponding image (which may be an image of a designated document, an image of the target user, or an image of a designated document and an image of the target user), collect the above image, and use the above image as the image taken of the target object.

S204: When the image does not meet the preset image quality conditions, the image is input into a pre-trained cause generation model, the cause generation model is used to determine the position of the image that does not meet the preset image quality conditions, and generate image reshoot guidance information for the determined position. The cause generation model is a model for detecting the reasons why the image does not meet the preset image quality conditions and providing corresponding improvement suggestions for the detection results.

S206: Re-shooting the image of the target object according to the image re-shooting guidance information.

The specific processing process of the above steps S204-S206 can refer to the relevant content of the above steps S104-S106, which will not be repeated here.

S208: If the image meets the preset image quality condition, the image and/or the information contained in the image is stored in the blockchain system.

When an image meets the preset image quality conditions, by storing the image and/or the information contained in the image in the blockchain system, the image information capacity in the blockchain system can be increased while ensuring the security and reliability of the image information.

Further, as shown in FIG3 , there may be a variety of methods for determining whether an image captured of a target object meets preset image quality conditions in the embodiments of this specification. Determining whether the captured image meets preset image quality conditions through a quality model is an optional processing method. When determining whether the captured image meets preset image quality conditions through a quality model, the image acquisition method of the embodiments of this specification may include the following steps S302-S308.

Step S302: Acquire an image captured of the target object.

The specific processing process of step S302 can refer to the relevant content of the above step S102, which will not be repeated here.

Step S304: input the image into a pre-trained quality model to obtain a quality assessment result of the image.

The quality assessment results of the image include: unqualified quality (i.e., the image does not meet the preset image quality conditions) and qualified quality (i.e., the image meets the preset image quality conditions). The quality model is used to determine whether the image meets the preset image quality conditions based on the image classification method, and the quality model is obtained by model training through image samples of different photographed objects and a preset second loss function. The input data of the quality model are image samples of different photographed objects, i.e., the images obtained in the above step S302, and the output results of the quality model are images with quality assessment results, including qualified quality images and unqualified quality images.

By training a quality model to obtain the quality assessment results of the image, it is possible to more quickly and accurately determine whether the current image meets the preset image quality conditions, thereby improving image acquisition efficiency.

In one embodiment, the quality model can be constructed using a binary classification algorithm, and the images taken for the target object can be divided into qualified quality images (or high-quality images) and unqualified quality images (or low-quality images) through the quality model.

In one embodiment, the second loss function may adopt a classification model loss function, specifically a cross entropy loss function, or a BCE-loss (binary cross entropy loss function).

If the quality assessment result indicates that the image meets the preset image quality condition, the user continues to perform subsequent processes based on the acquired image.

Step S306: If the quality assessment result indicates that the image does not meet the preset image quality conditions, the image is input into a pre-trained cause generation model, and the cause generation model is used to determine the position in the image that does not meet the preset image quality conditions, and generate image reshoot guidance information for the determined position. The cause generation model is a model for detecting the reasons why the image does not meet the preset image quality conditions, and providing corresponding improvement suggestions for the detection results.

Step S308: re-shooting the image of the target object according to the image re-shooting guidance information.

The specific processing process of the above steps S306-S308 can refer to the relevant content of the above steps S104-S106, which will not be repeated here.

Exemplarily, FIG4 takes the eKYC process as an example and shows a schematic principle diagram of an image acquisition method according to an embodiment of this specification. As can be seen from FIG4, the process of an image acquisition method provided by one or more embodiments of this specification in actual application is as follows.

1) In the eKYC process, the user completes the photo taking of ID or face.

2) After obtaining the image taken by the user, a pre-trained quality model based on image classification is used to perform a binary quality classification on the image taken by the user to determine whether the image taken by the user is a high-quality image or a low-quality image.

3) If the classification result is a high-quality image, the user continues to execute the subsequent process; if the classification result is a low-quality image, the image is input into a pre-trained cause generation model.

4) In the cause generation model, the causes of quality problems in low-quality images are subdivided. The causes of quality problems may include: highlights, low light, blur, etc. At the same time, the specific locations of quality problems in low-quality images are located in the cause generation model.

5) After obtaining the segmented causes and specific locations of the quality problems, text generation is performed through the cause generation model to obtain image reshoot guidance information, such as "the highlight in the upper left corner of the document is overexposed", "the overall image is blurred, please keep the phone still when shooting", etc., and the above image reshoot guidance information is fed back to the user.

6) The user retakes the image according to the obtained image retake guidance information and continues the eKYC process.

The embodiment of the present specification provides an image acquisition method, firstly acquiring an image shot for a target object, then determining whether the image meets a preset image quality condition, and when the image does not meet the preset image quality condition, inputting the image into a pre-trained cause generation model, determining the position in the image that does not meet the preset image quality condition through the cause generation model, and generating image reshoot guidance information for the determined position, and finally reshooting the target object according to the generated image reshoot guidance information, thereby completing image acquisition. When the image does not meet the preset image quality condition, the pre-trained cause generation model can detect the cause of the image not meeting the preset image quality condition in the last shooting process, obtain the specific cause of the quality problem, and locate the position in the last shot image that does not meet the preset image quality condition, obtain the specific position that causes the quality problem, and finally generate image reshoot guidance information for the determined position to feedback to the user, thereby providing targeted improvement suggestions for the next shooting, which can effectively reduce the number of repeated shootings, improve image shooting efficiency, and thus improve user experience.

In summary, specific embodiments of the present subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recorded in the claims can be performed in a different order and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing can be advantageous.

The above is an image acquisition method provided by one or more embodiments of this specification. Based on the same idea, one or more embodiments of this specification also provide an image acquisition device, as shown in FIG5 .

The image acquisition device includes: an image acquisition module 410 , an image reshooting guidance information generation module 420 and a control module 430 .

The image acquisition module 410 acquires an image taken of a target object.

The image reshooting guidance information generation module 420, when the image does not meet the preset image quality conditions, inputs the image into a pre-trained cause generation model, determines the position in the image that does not meet the preset image quality conditions through the cause generation model, and generates image reshooting guidance information for the determined position. The cause generation model is a model for detecting the cause of the image not meeting the preset image quality conditions and providing corresponding improvement suggestions for the detection results.

The control module 430 re-takes an image of the target object according to the image re-shooting guidance information.

In one embodiment, the image acquisition module 410 is also used to collect images of the target user's designated certificate and/or the target user taken by the target user to access the blockchain system based on electronic identity authentication information, and use the images of the designated certificate and/or the target user taken as the images taken of the target object; the image acquisition device may also include: a storage module, if the image meets the preset image quality conditions, the image and/or the information contained in the image is stored in the blockchain system.

In one embodiment, the image acquisition device may also include: a quality assessment result acquisition module, which inputs the image into a pre-trained quality model to obtain a quality assessment result of the image; an image reshooting guidance information generation module 420, which is also used to input the image into a pre-trained cause generation model if the quality assessment result indicates that the image does not meet the preset image quality conditions, determine the position in the image that does not meet the preset image quality conditions through the cause generation model, and generate image reshooting guidance information for the determined position.

The embodiment of the present specification provides an image acquisition device, which first acquires an image shot for a target object through an image acquisition module, and then, when the image does not meet the preset image quality conditions, the image retake guidance information generation module inputs the image into a pre-trained cause generation model, determines the position in the image that does not meet the preset image quality conditions through the cause generation model, and generates image retake guidance information for the determined position, and finally controls the image retake guidance information generation module to generate the image retake guidance information for the determined position. The module re-takes the image of the target object according to the generated image retake guidance information, thereby completing the image acquisition. When the image does not meet the preset image quality conditions, the pre-trained cause generation model can detect the reasons why the image did not meet the preset image quality conditions during the last shooting process, obtain the specific reasons for the quality problem, and locate the position in the last shot that does not meet the preset image quality conditions, obtain the specific position that causes the quality problem, and finally generate image retake guidance information for the determined position and feedback it to the user, so as to provide targeted improvement suggestions for the next shooting, which can effectively reduce the number of repeated shooting, improve image shooting efficiency, and thus improve user experience.

Those skilled in the art should understand that the above-mentioned image acquisition device can be used to implement the image acquisition method described above, and the detailed description thereof should be similar to the description of the method part above, and will not be further described here to avoid tediousness.

Based on the same idea, one or more embodiments of the present specification also provide an electronic device, as shown in FIG6 . The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors 501 and a memory 502, and the memory 502 may store one or more storage applications or data. Among them, the memory 502 may be a short-term storage or a persistent storage. The application stored in the memory 502 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the electronic device. Furthermore, the processor 501 may be configured to communicate with the memory 502 to execute a series of computer executable instructions in the memory 502 on the electronic device. The electronic device may also include one or more power supplies 503, one or more wired or wireless network interfaces 504, one or more input and output interfaces 505, and one or more keyboards 506.

Specifically in this embodiment, the electronic device includes a memory, and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions in the electronic device, and is configured to be executed by one or more processors. The one or more programs include the following computer executable instructions: obtaining an image taken of a target object; when the image does not meet a preset image quality condition, inputting the image into a pre-trained cause generation model, determining the position in the image that does not meet the preset image quality condition through the cause generation model, and generating image reshooting guidance information for the determined position, the cause generation model is a model for detecting the cause of the image not meeting the preset image quality condition, and providing corresponding improvement suggestions for the detection result; and re-taking the image of the target object according to the image reshooting guidance information.

One or more embodiments of the present specification also propose a storage medium for storing computer executable instructions, which implement the following process when executed by a processor: obtaining an image taken of a target object; when the image does not meet the preset image quality conditions, inputting the image into a pre-trained cause generation model; The cause generation model is used to determine the position in the image that does not meet the preset image quality conditions, and generate image reshoot guidance information for the determined position. The cause generation model is a model for detecting the reasons why the image does not meet the preset image quality conditions, and providing corresponding improvement suggestions for the detection results; according to the image reshoot guidance information, the target object is re-imaged.

The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For the convenience of description, the above devices are described in terms of functions and are divided into various units. Of course, when implementing one or more embodiments of this specification, the functions of each unit can be implemented in the same or multiple software and/or hardware.

Those skilled in the art will appreciate that one or more embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, one or more embodiments of this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, one or more embodiments of this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

One or more embodiments of the present specification are described with reference to the flowcharts and/or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present specification. It should be understood that each process and/or box in the flowchart and/or block diagram, as well as the combination of the processes and/or boxes in the flowchart and/or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the functions specified in one process or multiple processes in the flowchart and/or one box or multiple boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions can also be loaded into a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby performing a computer-implemented process on the computer or other programmable device. The instructions executed on other programmable devices provide steps for implementing the functions specified in one or more flows of the flowcharts and/or one or more blocks of the block diagrams.

In a typical configuration, a computing device includes one or more processors (CPU), input/output interfaces, network interfaces, and memory.

Memory may include non-permanent storage in a computer-readable medium, in the form of random access memory (RAM) and/or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined in this article, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

One or more embodiments of the present specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For the relevant parts, refer to the method embodiment. A partial description of the embodiments will suffice.

The above description is only one or more embodiments of this specification and is not intended to limit this application. For those skilled in the art, one or more embodiments of this specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of one or more embodiments of this specification should be included in the scope of the claims of one or more embodiments of this specification.

Claims (10)

An image acquisition method, comprising: Acquire an image taken of a target object; When the image does not meet the preset image quality condition, the image is input into a pre-trained cause generation model, the position of the image that does not meet the preset image quality condition is determined by the cause generation model, and image reshoot guidance information for the determined position is generated, wherein the cause generation model is a model for detecting the cause of the image not meeting the preset image quality condition and providing corresponding improvement suggestions for the detection result; The target object is imaged again according to the image reshooting guidance information. According to the method of claim 1, the training method of the cause generation model comprises: Acquire multiple image samples taken for different photographed objects, the reason why each image sample does not meet the preset image quality condition, and the position of each image sample that does not meet the preset image quality condition; Based on the multiple image samples, the reason why each image sample does not meet the preset image quality conditions, and the position in each image sample that does not meet the preset image quality conditions, the cause generation model is trained using a preset first loss function to obtain a trained cause generation model. According to the method of claim 2, the step of obtaining a plurality of image samples taken for different photographed objects, the reason why each image sample does not meet the preset image quality condition, and the position in each image sample that does not meet the preset image quality condition comprises: Acquire multiple image samples taken for different objects; The reason why each image sample does not meet the preset image quality condition and the position of each image sample that does not meet the preset image quality condition are determined respectively. According to the method of claim 2, the reason why the image sample does not meet the preset image quality condition includes: a first quality problem and a second quality problem, wherein the first quality problem is an image quality problem in which the image granularity is greater than or equal to a preset image granularity threshold, and the second quality problem is an image quality problem in which the image granularity is less than the preset image granularity threshold; Determining the position of each image sample that does not meet the preset image quality condition includes: When the reason why the image samples do not meet the preset image quality condition is a first quality problem, determining a position in each image sample that does not meet the preset image quality condition based on a target detection algorithm; When the reason why the image samples do not meet the preset image quality condition is the second quality problem, a position in each image sample that does not meet the preset image quality condition is identified based on a semantic segmentation algorithm. According to the method of claim 1, the acquiring of the image taken of the target object comprises: The target user is the target user photographed by the blockchain system based on electronic identity authentication information. The image of the designated certificate and/or the image of the target user is used as the image taken of the target object; The method further comprises: If the image meets the preset image quality condition, the image and/or the information contained in the image is stored in the blockchain system. According to the method according to claim 2, the preset first loss function adopts a cross entropy loss function. The method according to claim 1, further comprising: Inputting the image into a pre-trained quality model to obtain a quality assessment result of the image; When the image does not meet the preset image quality condition, the image is input into a pre-trained cause generation model, a position in the image that does not meet the preset image quality condition is determined by the cause generation model, and image reshooting guidance information for the determined position is generated, including: If the quality assessment result indicates that the image does not meet the preset image quality conditions, the image is input into a pre-trained cause generation model, the position of the image that does not meet the preset image quality conditions is determined by the cause generation model, and image reshoot guidance information is generated for the determined position. According to the method according to claim 7, the quality model is a model constructed using a binary classification algorithm. An image acquisition device, comprising: An image acquisition module, which acquires images taken of a target object; an image retake guidance information generating module, which, when the image does not meet the preset image quality condition, inputs the image into a pre-trained cause generation model, determines the position of the image that does not meet the preset image quality condition through the cause generation model, and generates image retake guidance information for the determined position, wherein the cause generation model is a model for detecting the cause causing the image to not meet the preset image quality condition and providing corresponding improvement suggestions for the detection result; A control module retakes an image of the target object according to the image retake guidance information. An electronic device, comprising: Processor; and a memory arranged to store computer executable instructions which, when executed, cause the processor to: Acquire an image taken of a target object; When the image does not meet the preset image quality condition, the image is input into a pre-trained cause generation model, the position of the image that does not meet the preset image quality condition is determined by the cause generation model, and image reshoot guidance information for the determined position is generated, wherein the cause generation model is used to determine the cause of the image not meeting the preset image quality condition. A model that detects the causes of preset image quality conditions and provides corresponding improvement suggestions for the detection results; The target object is imaged again according to the image reshooting guidance information.