Image database updating method, device, electronic device and medium

By replacing low-quality stored images into high-quality input images in the image database, the recognition verification failure caused by low-quality images in the image database is solved, and the accuracy and user experience of the recognition system are improved.

CN112148907BInactive Publication Date: 2024-11-19BEIJING BAIDU NETCOM SCI & TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011152109.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-10-23
Publication Date
2024-11-19
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The low quality of stored images in the image database causes image recognition verification to fail, reducing user experience.

Method used

By acquiring the similarity between the input image and the stored images in the database, determining the target image, and when the image quality evaluation is higher than the target image, replacing the target image with the input image to update the database and improving the image quality.

Benefits of technology

Improve the accuracy and user experience of image recognition verification, and enhance the reliability of the recognition system by updating the image quality in the database in real time.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112148907B_ABST
    Figure CN112148907B_ABST
Patent Text Reader

Abstract

The present application discloses an image database update method, device, electronic device and medium, and relates to the fields of deep learning, computer vision, etc. The image database update method includes: obtaining an input image; determining the similarity between the input image and multiple stored images in the database; determining a target image for the input image from multiple stored images based on the similarity; determining that the input image passes the recognition verification when the similarity between the input image and the target image is greater than a first similarity threshold; determining an image quality evaluation of the input image when the similarity between the input image and the target image is greater than a second similarity threshold, wherein the second similarity threshold is greater than the first similarity threshold; and replacing the target image with the input image to update the database when the image quality evaluation of the input image is higher than the image quality evaluation of the target image.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of computer technology, specifically, to the fields of deep learning and computer vision, and more specifically, to a method, device, electronic device and medium for updating an image database. Background Art

[0002] Image recognition technology is widely used. In the image recognition process, the current input image is usually obtained and the similarity between the input image and the stored image in the image database is compared. If the similarity reaches the similarity threshold, it can be determined that the input image has passed the verification. Taking the application of image recognition technology in payment scenarios as an example, the user uploads the face image to the image database in advance. When performing the payment operation later, the user's current face image is obtained, and the current face image is compared with the face image in the image database for similarity. If the similarity reaches the similarity threshold, it is determined that the face recognition has passed the verification and the user is allowed to pay. However, when the image quality of the stored image in the image database is poor, the image recognition verification will fail, thereby reducing the user's experience. Summary of the invention

[0003] The present application provides a method, device, electronic device and storage medium for updating an image database.

[0004] According to a first aspect, the present application provides a method for updating an image database, comprising: obtaining an input image, determining the similarity between the input image and a plurality of stored images in a database, determining a target image for the input image from the plurality of stored images based on the similarity, determining that the input image passes recognition verification when the similarity between the input image and the target image is greater than a first similarity threshold, determining an image quality evaluation of the input image when the similarity between the input image and the target image is greater than a second similarity threshold, wherein the second similarity threshold is greater than the first similarity threshold, and replacing the target image with the input image to update the database when the image quality evaluation of the input image is higher than the image quality evaluation of the target image.

[0005] According to a second aspect, the present application provides an image database update device, comprising: an acquisition module, a similarity determination module, a target image determination module, an identification determination module, an image quality evaluation determination module and an update module. The acquisition module is used to acquire an input image. The similarity determination module is used to determine the similarity between the input image and a plurality of stored images in a database. The target image determination module is used to determine a target image for the input image from the plurality of stored images based on the similarity. The identification determination module is used to determine that the input image passes the recognition verification when the similarity between the input image and the target image is greater than a first similarity threshold. The image quality evaluation determination module is used to determine the image quality evaluation of the input image when the similarity between the input image and the target image is greater than a second similarity threshold, wherein the second similarity threshold is greater than the first similarity threshold. The update module is used to replace the target image with the input image to update the database when the image quality evaluation of the input image is higher than the image quality evaluation of the target image.

[0006] According to a third aspect, the present application provides an electronic device, comprising: at least one processor and a memory in communication with the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the method as described above.

[0007] According to a fourth aspect, the present application provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the method as described above.

[0008] According to a fifth aspect, a computer program product is provided, comprising a computer program, which implements the above method when executed by a processor.

[0009] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present application.

[0011] Figure 1 The application scenario of image database update according to the embodiment of the present application is schematically shown;

[0012] Figure 2 A flowchart schematically shows a method for updating an image database according to an embodiment of the present application;

[0013] Figure 3 A flowchart schematically shows a method for updating an image database according to another embodiment of the present application;

[0014] Figure 4 A schematic diagram schematically shows an update of an image database according to an embodiment of the present application;

[0015] Figure 5 A flowchart schematically shows a method for updating an image database according to another embodiment of the present application;

[0016] Figure 6 A flowchart schematically shows a method for updating an image database according to another embodiment of the present application;

[0017] Figure 7 A flowchart of a training quality evaluation model according to an embodiment of the present application is schematically shown;

[0018] Figure 8 A schematic diagram of a training quality evaluation model according to an embodiment of the present application is schematically shown;

[0019] Fig. 9 A schematic diagram of a training quality evaluation model according to another embodiment of the present application is schematically shown;

[0020] Fig.10 A block diagram schematically shows an image database updating device according to an embodiment of the present application; and

[0021] Fig.11 It is a block diagram of an electronic device used to implement the image database updating method of the embodiment of the present application. DETAILED DESCRIPTION

[0022] The following is a description of exemplary embodiments of the present application in conjunction with the accompanying drawings, including various details of the embodiments of the present application to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0023] The terms used herein are only for describing specific embodiments and are not intended to limit the present application. The terms "include", "comprising", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0024] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0025] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0026] An embodiment of the present application provides a method for updating an image database, comprising: obtaining an input image, and determining the similarity between the input image and a plurality of stored images in the database. Then, based on the similarity, a target image for the input image is determined from the plurality of stored images, and when the similarity between the input image and the target image is greater than a first similarity threshold, it is determined that the input image has passed the recognition verification. Next, when the similarity between the input image and the target image is greater than a second similarity threshold, an image quality evaluation of the input image is determined, wherein the second similarity threshold is greater than the first similarity threshold, and when the image quality evaluation of the input image is higher than the image quality evaluation of the target image, the target image is replaced with the input image to update the database.

[0027] Figure 1 The application scenario of image database update according to the embodiment of the present application is schematically illustrated.

[0028] like Figure 1 As shown, the application scenario 100 of the embodiment of the present application includes, for example, a user 110 , a collection device 120 , and an image database 130 .

[0029] In the embodiment of the present application, the input image 111 of the user 110 may be collected by the collection device 120. The input image 111 of the user may be a face image of the user. After the collection device 120 collects the input image 111 of the user, image recognition may be performed on the input image 111.

[0030] The image recognition process can be performed by an electronic device. The electronic device can be a computer, a mobile phone, a server, etc. The acquisition device 120 can be installed in the electronic device, and the electronic device acquires the user's input image 111 through the acquisition device 120 and performs image recognition processing on the input image 111. Alternatively, the acquisition device 120 can be an external device independent of the electronic device, and the acquisition device 120 is data-connected to the electronic device. The acquisition device 120 can send the acquired input image 111 to the electronic device, so that the electronic device can perform image recognition processing on the received input image 111.

[0031] During the image recognition process, the input image 111 and the stored images in the image database 130 may be compared for similarity, and the stored images may include, for example, multiple images 131, 132, 133, etc. The stored images in the image database 130 may be pre-registered images, which may include facial images of multiple users. When the similarity between the input image 111 and at least one stored image in the image database 130 reaches a similarity threshold, it may be determined that the input image 111 has passed the recognition verification, and the user 110 is allowed to perform related operations. For example, taking the payment scenario as an example, when the user's input image 111 passes the recognition verification, the user 110 may be allowed to perform the payment operation.

[0032] The present application embodiment provides a method for updating an image database. Figure 1 For application scenarios, refer to Figure 2 to Figure 9 A method for updating an image database according to an exemplary embodiment of the present application will be described.

[0033] Figure 2 The flowchart schematically shows a method for updating an image database according to an embodiment of the present application.

[0034] like Figure 2 As shown, the image database updating method 200 of the embodiment of the present application may include, for example, operations S201 to S206.

[0035] In operation S201, an input image is acquired.

[0036] In operation S202, the similarity between the input image and a plurality of stored images in the database is determined.

[0037] In operation S203, a target image for the input image is determined from a plurality of stored images based on the similarity.

[0038] In the embodiment of the present application, the input image may be an image of a user currently captured, and the target image determined for the input image may be one of the multiple stored images. For example, when the multiple stored images include images of multiple users, the determined target image and the input image may be images belonging to the same user.

[0039] In operation S204, when the similarity between the input image and the target image is greater than a first similarity threshold, it is determined that the input image passes the recognition verification.

[0040] According to an embodiment of the present application, after determining the target image, the similarity between the input image and the target image can be compared, and it can be determined whether the similarity between the input image and the target image is greater than a first similarity threshold. The first similarity threshold can be a pre-set threshold, for example, the first similarity threshold can be 80%, 85%, and so on. In the case where the similarity between the input image and the target image is greater than the first similarity threshold, it can be determined that the input image passes the recognition verification. Taking the payment scenario as an example, when the input image passes the recognition verification, the user can be allowed to perform a payment operation. Those skilled in the art can understand that a variety of methods can be used to determine the similarity between the input image and the target image. In addition, those skilled in the art can understand that before calculating the similarity between the input image and the target image, the input image can be pre-processed, such as normalization, and the embodiments of the present application are not limited to this.

[0041] In operation S205, in a case where the similarity between the input image and the target image is greater than a second similarity threshold, an image quality evaluation of the input image is determined, and the second similarity threshold is greater than the first similarity threshold.

[0042] In operation S206, in a case where the image quality evaluation of the input image is higher than the image quality evaluation of the target image, the target image is replaced with the input image to update the database.

[0043] In an embodiment of the present application, when the similarity between the input image and the target image is greater than a first similarity threshold, it can be further determined whether the similarity between the input image and the target image is greater than a second similarity threshold. The second similarity threshold may be a pre-set threshold. The second similarity threshold is, for example, greater than the first similarity threshold. For example, the second similarity threshold may be 90%, 95%, and so on.

[0044] When the similarity between the input image and the target image is greater than the second similarity threshold, the image quality evaluation of the input image can be further determined, and the image quality evaluation of the input image and the image quality evaluation of the target image can be compared. If the image quality evaluation of the input image is higher than the image quality evaluation of the target image, the target image can be replaced with the input image, for example, the target image in the database is deleted, and the input image is stored in the database.

[0045] In the embodiments of the present application, the image quality evaluation, for example, characterizes the quality of the image. The image quality is related to factors such as the resolution and blur of the image. The image quality evaluation can be expressed as an evaluation score. The higher the evaluation score, the higher the image quality evaluation, that is, the higher the image quality. For example, when the image resolution is higher and the blur is lower, the image quality evaluation is higher. When the image resolution is lower and the blur is higher, the image quality evaluation is lower.

[0046] In the embodiment of the present application, by comparing the similarity between the input image and the target image, when the similarity is greater than the first similarity threshold, it is determined that the input image has passed the recognition verification, and when the similarity is greater than the second similarity threshold, it can be determined whether to update the database based on the image quality evaluation of the input image and the image quality evaluation of the target image. When the image quality evaluation of the input image is greater than the image quality evaluation of the target image, the input image can be used to replace the target image and stored in the database, so as to realize real-time updating of the database, improve the image quality of the images in the database, and thus improve the recognition accuracy of image recognition verification.

[0047] Figure 3 The following schematically shows a flow chart of a method for updating an image database according to another embodiment of the present application.

[0048] like Figure 3 As shown, the image database updating method 300 of the embodiment of the present application may include, for example, operations S301 to S311.

[0049] In operation S301, an input image is acquired.

[0050] In operation S302, the similarity between the input image and a plurality of stored images in the database is determined.

[0051] In operation S303, a target image for the input image is determined from a plurality of stored images based on the similarity.

[0052] In operation S304, when the similarity between the input image and the target image is greater than a first similarity threshold, it is determined that the input image passes the recognition verification.

[0053] In operation S305 , in a case where the similarity between the input image and the target image is greater than a second similarity threshold, an image quality evaluation of the input image is determined, and the second similarity threshold is greater than the first similarity threshold.

[0054] In operation S306, in a case where the image quality evaluation of the input image is higher than the image quality evaluation of the target image, the target image is replaced with the input image to update the database.

[0055] After operation S303 is performed, operation S307 may be performed.

[0056] In operation S307, it is determined whether the similarity between the input image and the target image is greater than a first similarity threshold. If yes, operation S304 is performed; otherwise, operation S308 is performed.

[0057] In operation S308, it is identified that the verification has failed.

[0058] After operation S304 is performed, operation S309 may be performed.

[0059] In operation S309, it is determined whether the similarity between the input image and the target image is greater than a second similarity threshold. If yes, operation S305 is performed; otherwise, operation S310 is performed.

[0060] In operation S310, the database is not updated.

[0061] After operation S305 is performed, operation S311 may be performed.

[0062] In operation S311, it is determined whether the image quality evaluation of the input image is higher than the image quality evaluation of the target image. If yes, operation S306 is performed; otherwise, operation S310 is performed.

[0063] In an embodiment of the present application, the difference between the second similarity threshold and the first similarity threshold is set to be less than a preset difference, so as to ensure that the second similarity threshold is not too high. Wherein, the difference between the second similarity threshold and the first similarity threshold is less than a preset difference, for example, 10%, which can avoid that when the input image is used to replace the target image, the change between the updated database obtained after the replacement and the database before the update is small due to the large similarity between the input image and the target image, resulting in poor database update effect. That is, if the similarity between the input image and the target image is too large, it will be difficult to substantially improve the image quality of the images stored in the database by replacing the target image with the input image to update the database. Therefore, the embodiment of the present application improves the update effect of the database by setting the difference between the second similarity threshold and the first similarity threshold to be less than the preset difference.

[0064] Figure 4 The figure schematically shows the updating of the image database according to the embodiment of the present application.

[0065] like Figure 4 As shown, description is made by taking an input image 411 and a database 420 having three stored images 421 , 422 , and 423 as an example.

[0066] The input image 411 is compared with each stored image to obtain the similarity between the input image and the multiple stored images 421, 422, and 423. The stored image with the greatest similarity is used as the target image. For example, since the similarity between the input image 411 and the stored image 421 is the highest, the stored image 421 is determined to be the target image. When the similarity between the input image 411 and the stored image 421 is greater than a first similarity threshold, such as 85%, it is determined that the input image 411 passes the recognition verification. Then, it is determined whether the similarity between the input image 411 and the stored image 421 is greater than a second similarity threshold, such as 90%. If the similarity between the input image 411 and the stored image 421 is greater than the second similarity threshold, the stored image 421 is replaced with the input image 411, for example, the stored image 421 is deleted from the database 420, and the input image 411 is stored in the database 420.

[0067] Figure 5 The following schematically shows a flow chart of a method for updating an image database according to another embodiment of the present application.

[0068] like Figure 5 As shown, the image database updating method 500 of the embodiment of the present application may include, for example, operations S501 to S512.

[0069] In operation S501, an input image is acquired.

[0070] In operation S502, the similarity between the input image and a plurality of stored images in the database is determined, for example, by processing the input image and the plurality of stored images using an image recognition model to determine the similarity between the input image and the plurality of stored images.

[0071] In an embodiment of the present application, an input image and a plurality of stored images can be processed using an image recognition model to obtain image features of the input image and image features of each stored image, and then based on the image features of the input image and the image features of each stored image, the similarity between the input image and each stored image is determined. The image features can be represented as feature vectors of the image, and the similarity is determined by calculating the distance between the feature vectors of the input image and the feature vectors of the stored images, wherein the smaller the distance, the higher the similarity, and the larger the distance, the lower the similarity. It can be understood by those skilled in the art that the distance between feature vectors includes but is not limited to Euclidean distance, angle cosine distance, and Manhattan distance, and the embodiment of the present application does not limit this.

[0072] In an embodiment of the present application, the image recognition model may include but is not limited to a convolutional neural network model.

[0073] In operation S503, a target image for the input image is determined from the plurality of stored images based on the similarity. For example, a stored image with the greatest similarity to the input image among the plurality of stored images is determined as the target image.

[0074] In operation S504, when the similarity between the input image and the target image is greater than a first similarity threshold, it is determined that the input image passes the recognition verification.

[0075] In operation S505 , when the similarity between the input image and the target image is greater than a second similarity threshold, an image quality evaluation of the input image is determined using a trained quality evaluation model, and the second similarity threshold is greater than the first similarity threshold.

[0076] In an embodiment of the present application, a trained quality assessment model may be used to perform quality assessment on an input image to obtain an image quality assessment of the input image. The quality assessment model may include but is not limited to a convolutional neural network model, and the input image is input into the convolutional neural network model for processing to obtain an image quality assessment of the input image. Performing quality assessment on an input image using a trained quality assessment model may improve the accuracy of the quality assessment.

[0077] In operation S506, in a case where the image quality evaluation of the input image is higher than the image quality evaluation of the target image, the target image is replaced with the input image to update the database.

[0078] After operation S503 is performed, operation S507 may be performed.

[0079] In operation S507, it is determined whether the similarity between the input image and the target image is greater than a first similarity threshold. If yes, operation S504 is performed; otherwise, operation S508 is performed.

[0080] In operation S508, it is identified that the verification has failed.

[0081] After operation S504 is performed, operation S509 may be performed.

[0082] In operation S509, it is determined whether the similarity between the input image and the target image is greater than a second similarity threshold. If yes, operation S505 is performed; otherwise, operation S510 is performed.

[0083] In operation S510, the database is not updated.

[0084] After operation S505 is performed, operation S511 may be performed.

[0085] In operation S511, it is determined whether the image quality evaluation of the input image is higher than the image quality evaluation of the target image. If yes, operation S506 is performed; otherwise, operation S510 is performed.

[0086] According to an embodiment of the present application, operation S512 may be performed before operation S501 is performed.

[0087] In operation S512, quality evaluation is performed on each of the plurality of stored images based on the quality evaluation model to obtain an image quality evaluation of each stored image.

[0088] In one example, an embodiment of the present application can perform image quality evaluation on stored images in a database in advance, and store the image quality evaluation for each stored image in association with the stored image in the database, so that after the target image is subsequently determined, the image quality evaluation of the input image can be compared with the image quality evaluation of the predetermined target image, thereby improving the comparison efficiency of the image quality evaluation.

[0089] In another example, an embodiment of the present application can use a quality evaluation model to process the input image and the target image in real time to obtain the image quality evaluation of the input image and the image quality evaluation of the target image when comparing the image quality evaluation of the input image and the image quality evaluation of the target image, so as to perform quality evaluation on the input image and the target image that need to be compared in a targeted manner.

[0090] Figure 6 The following schematically shows a flow chart of a method for updating an image database according to another embodiment of the present application.

[0091] like Figure 6 As shown, the image database updating method 600 of the embodiment of the present application may include, for example, operations S610 to S6110.

[0092] In operation S610, a database image is registered.

[0093] In operation S620, a plurality of stored images are obtained after the image registration.

[0094] In operation S630, an input image is acquired.

[0095] In operation S640, it is determined whether the similarity between the input image and the target image in the plurality of stored images is greater than a first similarity threshold, if yes, operation S650 is performed, if not, operation S670 is performed.

[0096] In operation S650, it is determined whether the similarity between the input image and the target image in the plurality of stored images is greater than a second similarity threshold, if yes, operation S660 is performed, if not, operation S670 is performed.

[0097] In operation S660, the stored image and the input image are input into a quality assessment model for quality assessment.

[0098] In operation S670, the database is not updated.

[0099] In operation S680, quality evaluation is performed on each stored image using a quality evaluation model to obtain an image quality evaluation of each stored image.

[0100] In operation S690, the quality of the input image is evaluated by using a quality evaluation model to obtain an image quality evaluation of the input image.

[0101] In operation S6100, it is determined whether the image quality evaluation of the input image is higher than the image quality evaluation of the target image. If yes, operation S6110 is performed, and if not, operation S670 is performed.

[0102] In operation S6110, the database is updated, ie, the target image is replaced with the input image.

[0103] Figure 7 The flowchart of the training quality evaluation model according to the embodiment of the present application is schematically shown.

[0104] like Figure 7 As shown, the training quality evaluation model of the embodiment of the present application may include, for example, operations S720 to S760.

[0105] In operation S720, the plurality of sample images are processed using an image recognition model to obtain similarities between a reference image in the plurality of sample images and other sample images except the reference image.

[0106] In operation S740, the image quality of the reference image is used as a reference, and an image quality evaluation of each of the other sample images is obtained based on the similarity between the other sample images and the reference image.

[0107] In operation S760, a quality assessment model is trained based on the plurality of sample images and the image quality assessment of each of the plurality of sample images.

[0108] In an embodiment of the present application, a sample image with the highest image quality can be designated as a reference image from a plurality of sample images. Then, the reference image and other sample images are processed using an image recognition model to obtain the similarity between the reference image and other sample images. Next, the image quality evaluation of each other sample image is determined based on the similarity between the other sample images and the reference image, taking the image quality of the reference image as a reference. The higher the similarity between the other sample images and the reference image, the higher the image quality evaluation of the sample image, that is, the closer the image quality evaluation of the sample image is to the image quality evaluation of the reference image.

[0109] After obtaining the image quality evaluation of each sample image, the image quality evaluation of each sample image can be used as the label information of the sample image, and the quality evaluation model can be trained based on the sample images with label information, so that the image quality evaluation output by the quality evaluation model when processing the sample image is closer to the image quality evaluation indicated by the label information of the sample image.

[0110] In the embodiment of the present application, the sample image is processed by using the image recognition model to obtain the image quality evaluation of the sample image, thereby improving the accuracy of the image quality evaluation of the sample image. The quality evaluation model is trained based on the sample image and its corresponding image quality evaluation, thereby improving the evaluation accuracy of the quality evaluation model, so that the image quality evaluation obtained by performing quality evaluation on the input image and the images stored in the database based on the trained quality evaluation model is more accurate.

[0111] According to the embodiment of the present application, after the image recognition model is optimized and updated, the sample image can be reprocessed based on the updated image recognition model to obtain the image quality evaluation of the updated sample image. Then, the quality evaluation model is retrained based on the sample image and its updated image quality evaluation, so as to facilitate the subsequent re-evaluation of the quality of the input image and multiple stored images based on the updated quality evaluation model, thereby optimizing the quality evaluation process to improve the update effect of the database.

[0112] Figure 8 A schematic diagram of a training quality evaluation model according to an embodiment of the present application is schematically shown.

[0113] like Figure 8As shown, the embodiment of the present application includes multiple sample images 811 , 812 , 813 , 814 , as well as an image recognition model 820 and a quality assessment model 830 .

[0114] For example, a sample image is designated as a reference image from a plurality of sample images 811, 812, 813, and 814, for example, sample image 811 is designated as a reference image, and an image quality evaluation of the reference image is designated, for example, the image quality evaluation of the reference image is designated as 100 points. The plurality of sample images 811, 812, 813, and 814 are input into the image recognition model 820 for processing, and the similarities between the sample images 812, 813, and 814 and the sample image 811 are obtained. For example, the similarity between the sample image 812 and the sample image 811 is 95%, the similarity between the sample image 813 and the sample image 811 is 90%, and the similarity between the sample image 814 and the sample image 811 is 85%.

[0115] After obtaining the similarity between sample images 812, 813, 814 and sample image 811, the image quality evaluation of each of sample images 812, 813, and 814 is obtained with reference to the image quality score of sample image 811. The greater the similarity between a sample image and sample image 811, the higher the image quality evaluation of the sample image, that is, the closer it is to the image quality evaluation of sample image 811. For example, the image quality evaluation of sample image 812 may be 95 points, the image quality evaluation of sample image 813 may be 90 points, and the image quality evaluation of sample image 814 may be 85 points. In one example, the evaluation score of the image quality evaluation of other sample images may be, for example, the similarity between the sample image and the reference image multiplied by the evaluation score of the image quality evaluation of the reference image. Taking sample image 812 as an example, the evaluation score of the image quality evaluation of the sample image 812 may be, for example, 95%*100 points=95 points.

[0116] After obtaining the image quality evaluations of the multiple sample images 811 , 812 , 813 , 814 , the multiple sample images 811 , 812 , 813 , 814 and their respective corresponding image quality evaluations may be input into the quality evaluation model 830 to train the quality evaluation model 830 .

[0117] Fig. 9 A schematic diagram of a training quality evaluation model according to another embodiment of the present application is schematically shown.

[0118] like Fig. 9As shown, the multiple sample images of the embodiment of the present application include, for example, multiple groups of sample images 910A, 910B, 910C, and include an image recognition model 920 and a quality assessment model 930. Among them, each group of sample images in the multiple groups of sample images 910A, 910B, and 910C is for one user. For example, the first group of images 910A includes multiple sample images 911A, 912A, 913A, and 914A of user 1, the second group of images 910B includes multiple sample images 911B, 912B, 913B, and 914B of user 2, and the third group of images 910C includes multiple sample images 911C, 912C, 913C, and 914C of user 3.

[0119] For each group of sample images in the multiple groups of sample images, each group of sample images includes a reference image and other sample images except the reference image.

[0120] Taking the first group of sample images 910A as an example, a sample image is designated as a reference image from a plurality of sample images 911A, 912A, 913A, and 914A, for example, the sample image 911A is designated as the reference image, and the image quality evaluation of the reference image is designated, for example, the image quality evaluation of the reference image is 100 points. The plurality of sample images 911A, 912A, 913A, and 914A are input into the image recognition model 920 for processing, and the similarities between the sample images 912A, 913A, and 914A and the sample image 911A are obtained. For example, the similarity between the sample image 912A and the sample image 911A is 95%, the similarity between the sample image 913A and the sample image 911A is 90%, and the similarity between the sample image 914A and the sample image 911A is 85%.

[0121] After obtaining the similarity between sample images 912A, 913A, 914A and sample image 911A, the image quality evaluation of each of sample images 912A, 913A, and 914A is obtained with reference to the image quality score of sample image 911A. The greater the similarity between a sample image and sample image 911A, the higher the image quality evaluation of the sample image, that is, the closer it is to the image quality evaluation of sample image 911A. For example, the image quality evaluation of sample image 912A may be 95 points, the image quality evaluation of sample image 913A may be 90 points, and the image quality evaluation of sample image 914A may be 85 points. In one example, the evaluation score of the image quality evaluation of other sample images may be, for example, the similarity between the sample image and the reference image multiplied by the evaluation score of the image quality evaluation of the reference image. Taking sample image 912A as an example, the evaluation score of the image quality evaluation of sample image 912A may be, for example, 95%*100 points=95 points.

[0122] The same as or similar to the first group of sample images 910A, the image quality evaluation of each sample image in each group of sample images can be obtained. Fig. 9 The image quality evaluation of each sample image in the first group of sample images 910A is marked as an example. After obtaining the image quality evaluation of each sample image in the multiple groups of sample images, each sample image and its corresponding image quality evaluation can be input into the quality evaluation model 930 to train the quality evaluation model 930. The quality evaluation model is trained using multiple groups of sample data, thereby improving the evaluation accuracy of the quality evaluation model.

[0123] The image database updating method of the embodiment of the present application can be applied to the field of face recognition. For example, the input image includes a face image, the multiple stored images include multiple pre-stored face images, and the recognition verification of the input image includes face recognition verification of the face image.

[0124] Fig.10 The block diagram schematically shows a device for updating an image database according to an embodiment of the present application.

[0125] like Fig.10 As shown, the image database updating device 1000 of the embodiment of the present application includes, for example, an acquisition module 1001, a similarity determination module 1002, a target image determination module 1003, an identification determination module 1004, an image quality evaluation determination module 1005 and an update module 1006.

[0126] The acquisition module 1001 may be used to acquire an input image. According to an embodiment of the present application, the acquisition module 1001 may, for example, execute the above reference Figure 2 The operation S201 described above will not be described in detail here.

[0127] The similarity determination module 1002 may be used to determine the similarity between the input image and a plurality of stored images in the database. According to an embodiment of the present application, the similarity determination module 1002 may, for example, execute the above reference Figure 2 The operation S202 described above will not be described in detail here.

[0128] The target image determination module 1003 may be used to determine a target image for the input image from a plurality of stored images based on similarity. According to an embodiment of the present application, the target image determination module 1003 may, for example, execute the above reference Figure 2 The operation S203 described above will not be described in detail here.

[0129] The recognition determination module 1004 can be used to determine that the input image passes the recognition verification when the similarity between the input image and the target image is greater than the first similarity threshold. According to the embodiment of the present application, the recognition determination module 1004 can, for example, execute the above reference Figure 2 The operation S204 described above will not be described in detail here.

[0130] The image quality evaluation determination module 1005 may be configured to determine the image quality evaluation of the input image when the similarity between the input image and the target image is greater than a second similarity threshold, wherein the second similarity threshold is greater than the first similarity threshold. According to an embodiment of the present application, the image quality evaluation determination module 1005 may, for example, execute the above reference Figure 2 The operation S205 described above will not be described in detail here.

[0131] The updating module 1006 may be used to replace the target image with the input image to update the database when the image quality evaluation of the input image is higher than the image quality evaluation of the target image. Figure 2 The operation S206 described above will not be described in detail here.

[0132] According to an embodiment of the present application, the present application further provides an electronic device, a readable storage medium and a computer program product. The computer program product includes a computer program, and the computer program can implement the method of any of the above embodiments when executed by a processor.

[0133] Fig.11 It is a block diagram of an electronic device used to implement the image database updating method of the embodiment of the present application.

[0134] like Fig.11 , is a block diagram of an electronic device 1100 according to an update method of an image database of an embodiment of the present application. The electronic device 1100 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, 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, 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 application described and / or required herein.

[0135] like Fig.11As shown, the electronic device 1100 includes: one or more processors 1110, a memory 1120, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses, and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the electronic device 1100, including instructions stored in or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In other embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple electronic devices 1100 can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig.11 A processor 1110 is taken as an example.

[0136] The memory 1120 is a non-transient computer-readable storage medium provided in the present application. The memory stores instructions executable by at least one processor to enable at least one processor to perform the image database update method provided in the present application. The non-transient computer-readable storage medium of the present application stores computer instructions, which are used to enable a computer to perform the image database update method provided in the present application.

[0137] The memory 1120 is a non-transient computer-readable storage medium that can be used to store non-transient software programs, non-transient computer executable programs and modules, such as program instructions / modules corresponding to the image database update method in the embodiment of the present application (for example, the attached Fig.10 The processor 1110 executes various functional applications and data processing of the server by running the non-transient software programs, instructions and modules stored in the memory 1120, that is, the image database update method in the above method embodiment is implemented.

[0138] The memory 1120 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function; the data storage area may store data created by the use of the electronic device 1100 according to the update of the image database, etc. In addition, the memory 1120 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage devices. In some embodiments, the memory 1120 may optionally include a memory remotely arranged relative to the processor 1110, and these remote memories may be connected to the electronic device 1100 for updating the image database via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0139] The electronic device 1100 of the method for updating the image database may further include: an input device 1130 and an output device 1140. The processor 1110, the memory 1120, the input device 1130 and the output device 1140 may be connected via a bus or other means. Fig.11 The example of connecting through bus is taken in the following.

[0140] The input device 1130 can receive input digital or character information, and generate key signal input related to user settings and function control of the electronic device 1100 for updating the image database, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator rod, one or more mouse buttons, a trackball, a joystick, and other input devices. The output device 1140 may include a display device, an auxiliary lighting device (e.g., an LED), and a tactile feedback device (e.g., a vibration motor), etc. The display device may include, but is not limited to, a liquid crystal display (LCD), a light emitting diode (LED) display, and a plasma display. In some embodiments, the display device may be a touch screen.

[0141] Various implementations of the systems and techniques described herein can be realized in digital electronic circuit systems, integrated circuit systems, dedicated ASICs (application specific integrated circuits), 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.

[0142] These computer programs (also referred to as programs, software, software applications, or code) include machine instructions for programmable processors and can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or means (e.g., disk, optical disk, memory, programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0143] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer 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 pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0144] The systems and techniques described herein may be implemented in a computing system that includes back-end 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 front-end 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 back-end components, middleware components, or front-end 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), and the Internet.

[0145] A computer system may include clients and servers. Clients and servers are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship to each other.

[0146] 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 recorded in this application can be executed in parallel, sequentially or in different orders, as long as the expected results of the technical solution disclosed in this application can be achieved, and this document is not limited here.

[0147] The above specific implementations do not constitute a limitation on the protection scope of this application. 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 modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A method for updating an image database, comprising: Get the input image; Determining the similarity between the input image and a plurality of stored images in a database; Based on the similarity, determining a target image for the input image from the plurality of stored images; When the similarity between the input image and the target image is greater than a first similarity threshold, determining that the input image passes the recognition verification; determining an image quality evaluation of the input image when the similarity between the input image and the target image is greater than a second similarity threshold, wherein the second similarity threshold is greater than the first similarity threshold; and When the image quality evaluation of the input image is higher than the image quality evaluation of the target image, replacing the target image with the input image to update the database; The image quality evaluation of the input image is obtained by processing the input image using a trained quality evaluation model, and the trained quality evaluation model is trained by the following operations: Specifying a sample image with the highest image quality from a plurality of sample images as a reference image; Taking the image quality of the benchmark image as a reference, obtaining an image quality evaluation of each sample image in the other sample images based on the similarity between the other sample images and the benchmark image; and The quality assessment model is trained based on the multiple sample images and the image quality assessment of each of the multiple sample images.

2. The method according to claim 1, wherein: Determining the similarity between the input image and a plurality of stored images in a database comprises: Processing the input image and the plurality of stored images using an image recognition model to obtain image features of the input image and image features of each stored image; and Based on the image features of the input image and the image features of each stored image, a similarity between the input image and each stored image is determined.

3. The method according to claim 1, wherein: The similarity between the other sample images and the reference image is obtained by the following operation: The image recognition model is used to process a plurality of sample images to obtain similarities between a reference image in the plurality of sample images and other sample images except the reference image.

4. The method according to claim 1, wherein: The multiple sample images include multiple groups of sample images, each group of sample images is for one user; The reference image and the other sample images except the reference image belong to the same group of sample images among multiple groups of sample images.

5. The method according to claim 1, further comprising: A quality evaluation is performed on each of the plurality of stored images based on the quality evaluation model to obtain an image quality evaluation of each stored image.

6. The method according to claim 1, wherein: The determining, based on the similarity, a target image for the input image from the plurality of stored images comprises: A stored image having the greatest similarity to the input image among the plurality of stored images is determined as the target image.

7. The method according to claim 1, wherein: The difference between the second similarity threshold and the first similarity threshold is smaller than a preset difference.

8. The method according to claim 1, wherein: The input image includes a face image; the multiple stored images include multiple pre-stored face images; and performing recognition verification on the input image includes performing face recognition verification on the face image.

9. An image database updating device, comprising: An acquisition module, used for acquiring an input image; A similarity determination module, used to determine the similarity between the input image and a plurality of stored images in a database; a target image determination module, configured to determine a target image for the input image from the plurality of stored images based on the similarity; an identification determination module, for determining that the input image passes the identification verification if the similarity between the input image and the target image is greater than a first similarity threshold; an image quality evaluation determination module, configured to determine an image quality evaluation of the input image if the similarity between the input image and the target image is greater than a second similarity threshold, wherein the second similarity threshold is greater than the first similarity threshold; and An updating module, configured to replace the target image with the input image to update the database if the image quality evaluation of the input image is higher than the image quality evaluation of the target image; The image quality evaluation of the input image is obtained by processing the input image using a trained quality evaluation model, and the trained quality evaluation model is trained by the following operations: Specifying a sample image with the highest image quality from a plurality of sample images as a reference image; Taking the image quality of the benchmark image as a reference, obtaining an image quality evaluation of each sample image in the other sample images based on the similarity between the other sample images and the benchmark image; and The quality assessment model is trained based on the multiple sample images and the image quality assessment of each of the multiple sample images.

10. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

11. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 8.