CN113221918B - Object detection method, object detection model training method and device - Google Patents

Abstract

The disclosure provides a target detection method, a target detection model training method and a device, and relates to the technical field of artificial intelligence. The method includes: obtaining an image to be detected; inputting the image to be detected into a pre-trained target detection model to obtain a target detection result corresponding to the image to be detected; when the target detection result satisfies a preset condition, obtaining a position box in the target detection result information; wherein, the preset condition is that the target detection result contains location frame information, and the confidence corresponding to the target detection result is less than the preset confidence threshold; the target detection result is verified based on the location frame information, and the image corresponding to the image to be detected is obtained. The target verification result of . In the technical solution of the present disclosure, among the target detection results output by the target detection model, the target detection results containing the location frame information and whose confidence is less than the preset confidence threshold are verified, so as to improve the accuracy of image review and contain The recall rate of the image of the target object.

Description

Object detection method, object detection model training method and device

technical field

The present disclosure relates to the technical field of artificial intelligence, in particular to the field of machine learning.

Background technique

With the rise of e-commerce, business advertising methods are becoming more and more diverse, and rich image promotion is an important means of advertising. In the promotion process of actual advertising materials, there are cases where merchants publish images containing specific targets in violation of regulations in the process of graphic and text promotion, and it is necessary to review the pictures provided in the advertising materials of merchants. The accuracy rate of the content review scheme in the prior art is low , unable to meet the actual needs.

Contents of the invention

The disclosure provides a target detection method, a target detection model training method, a device, a device, and a storage medium.

According to an aspect of the present disclosure, a target detection method is provided, including:

Obtain the image to be detected;

Input the image to be detected into the pre-trained target detection model to obtain the target detection result corresponding to the image to be detected;

When the target detection result satisfies the preset condition, obtain the location frame information in the target detection result; where the preset condition is that the target detection result contains the location frame information, and the confidence corresponding to the target detection result is less than the preset confidence degree threshold;

The target detection result is verified based on the position frame information, and the target verification result corresponding to the image to be detected is obtained.

According to another aspect of the present disclosure, a method for training a target detection model is provided, including:

Obtain multiple sets of first training sample data, and train the initial target detection model based on the multiple sets of first training sample data until a preset training end condition is met, to obtain the target detection model of any embodiment of the present disclosure;

Wherein, the first training sample data includes the first sample image and the first sample label corresponding to the first sample image, and the first sample label is used to represent whether the first sample image contains the target object and the position of the target object frame information, the first sample image includes a mosaic image; the mosaic image is obtained by stitching based on the target object image and similar target images, the target object image is an image containing the target object, the similar target image is an image containing similar target objects, and the similar The similarity between the target object and the target object is within a preset range.

According to another aspect of the present disclosure, a target detection device is provided, including:

The detection image acquisition module is used to obtain the image to be detected;

The target detection module is used to input the image to be detected into the pre-trained target detection model to obtain the target detection result corresponding to the image to be detected;

The position frame acquisition module is used to obtain the position frame information in the target detection result when the target detection result meets the preset condition; wherein, the preset condition is that the target detection result contains the position frame information, and the target detection result corresponds to The confidence level is less than the preset confidence level threshold;

The result verification module is configured to verify the target detection result based on the location frame information, and obtain the target verification result corresponding to the image to be detected.

According to another aspect of the present disclosure, a training device for a target detection model is provided, including:

A sample acquisition module, configured to acquire multiple sets of first training sample data;

A model training module, configured to train an initial target detection model based on multiple sets of first training sample data, until a preset training end condition is met, to obtain the target detection model of any embodiment of the present disclosure;

Wherein, the first training sample data includes the first sample image and the first sample label corresponding to the first sample image, and the first sample label is used to represent whether the first sample image contains the target object and the position of the target object frame information, the first sample image includes a mosaic image; the mosaic image is obtained by stitching based on the target object image and similar target images, the target object image is an image containing the target object, the similar target image is an image containing similar target objects, and the similar The similarity between the target object and the target object is within a preset range.

According to another aspect of the present disclosure, an electronic device is provided, including:

at least one processor; and

a memory communicatively coupled to 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 execute the method in any embodiment of the present disclosure.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the method in any embodiment of the present disclosure.

According to another aspect of the present disclosure, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the method in any embodiment of the present disclosure is implemented.

The disclosed technical solution solves the problem of low review accuracy in the content review solution in the prior art. In the target detection method in the disclosed technical solution, in the target detection results output by the target detection model, the target detection results containing position frame information and whose confidence level is less than the preset confidence level threshold are verified, so as to improve the accuracy of image review. Precision and recall for images containing the target object.

It should be understood that what is described in this section is not intended to identify key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will be readily understood through the following description.

Description of drawings

The accompanying drawings are used to better understand the present solution, and do not constitute a limitation to the present disclosure. in:

FIG. 1 is a schematic diagram of a target detection method in an embodiment of the present disclosure;

2 is a schematic diagram of the process of obtaining training samples of a target detection model in an embodiment of the present disclosure;

FIG. 3 is a schematic diagram of an iterative process of a target detection model in an embodiment of the present disclosure;

FIG. 4 is a schematic diagram of an image review system in an embodiment of the present disclosure;

5 is a schematic diagram of a training method for a target detection model in an embodiment of the present disclosure;

6 is a schematic diagram of a target detection device in an embodiment of the present disclosure;

7 is a schematic diagram of a sample update module in an embodiment of the present disclosure;

8 is a schematic diagram of a training device for a target detection model in an embodiment of the present disclosure;

FIG. 9 is a block diagram of an electronic device used to implement the object detection method of the embodiment of the present disclosure.

Detailed ways

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and they should be regarded as exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

In related technologies, when performing target detection on content in an image, the following two methods are generally used:

One way is to use a sliding window method for target area extraction and target search. For multi-category target recognition, features are extracted based on histogram of oriented gradient (Histogram of Oriented Gradient, HOG), scale-invariant feature transform (Scale-invariant feature transform, SIFT) and other feature extraction methods, and training based on adaptive enhancement ( Adaptive boosting, Adaboost) algorithm and support vector machine (Support Vector Machine, SVM) and other traditional machine learning classifiers to determine whether the image contains the target. Using this method for target detection has poor robustness, low recognition accuracy and slow speed, and it is difficult to meet the requirements of content review delay and accuracy.

Another way is to detect targets based on the deep learning target detection scheme. By collecting a large amount of image data containing targets, and manually marking the detection position information of various targets, the trained model identifies the target area and records the position of the regression frame. Detections are determined by class thresholding. This target detection method relies on a large amount of labeled data for target detection model training, and the collection of labeled data is difficult. Through the collected data with a single background, the accuracy and recall of the trained model are low in scenarios where risk data is extremely sparse. , unable to meet the content review requirements of high recall and high accuracy. For certain categories of target objects that have few chances of daily occurrence, the data that can be collected cannot reach the level that the target detection model can train. In addition, the actual needs cannot be met for target objects that need to be exempted from auditing in specific scenarios.

In the technical solution of the present disclosure, among the target detection results output by the target detection model, the target detection results containing the location frame information and whose confidence is less than the preset confidence threshold are verified, so as to improve the accuracy of image review and contain The recall rate of the image of the target object. By splicing target object images and similar target object images, the spliced images obtained are used as training samples, which can enrich the training sample data of the model. Through the audit exemption mechanism, the audit exemption requirements of images to be detected in specific scenarios are met.

The subject of execution of the present disclosure may be any electronic device, for example, a server or the like. The target detection method in the embodiment of the present disclosure will be introduced in detail below.

FIG. 1 is a schematic diagram of a target detection method in an embodiment of the present disclosure. As shown in Figure 1, object detection methods can include:

Step S101, acquiring an image to be detected;

Step S102, input the image to be detected into the pre-trained target detection model, and obtain the target detection result corresponding to the image to be detected;

Step S103, when the target detection result satisfies the preset condition, acquire the location frame information in the target detection result; wherein, the preset condition is that the target detection result contains the location frame information, and the confidence degree corresponding to the target detection result is less than the preset set the confidence threshold.

In step S104, the target detection result is verified based on the position frame information, and the target verification result corresponding to the image to be detected is obtained.

In the target detection method of the embodiment of the present disclosure, in the target detection results output by the target detection model, the target detection results containing the location frame information and whose confidence is less than the preset confidence threshold are verified, so as to improve the accuracy of image review rate and recall of images containing the target object.

In an implementation manner, for example, in an application scenario of image review, the image to be detected may be an image requiring content review. The image may include any target object that cannot be approved, and the target object can include a badge image. In addition, the image to be detected may also be an image in other application scenarios requiring target detection, which is not limited in this application.

In one embodiment, the target detection model can be any neural network model, including but not limited to:

Target detection algorithms such as fast region-convolution neural network (FasterRCNN), Single Shot MultiBox Detector (SSD), Yolo (You Only LookOnce), etc.

In one embodiment, the target detection result output by the target detection model may include indication information whether the target object is included, for example, the number 0 indicates that the target object is not included, and the number 1 indicates that the target object is included. Other methods may also be used to indicate whether the image to be detected contains the target object, which is not limited in this application. If the target detection result contains the target object, output the location frame information of the target object in the image to be detected and the confidence corresponding to the target detection result. Wherein, the location frame information includes but not limited to coordinate information of the location frame and the like.

In one embodiment, the confidence threshold is pre-configured according to specific needs. If the confidence corresponding to the target detection result is less than the confidence threshold, the image corresponding to the position frame can be determined in the image to be detected according to the position frame information. , to perform verification, and further determine whether the target object is included in the image to be detected, so as to improve the accuracy and recall rate of target detection.

In one embodiment, the target detection model is obtained by using multiple sets of first training sample data to train the initial target detection model. The first training sample data includes the first sample image and the first sample image corresponding to the first sample image. Sample label, the first sample label is used to represent whether the first sample image contains the target object and the location frame information of the target object; the first sample image includes a stitched image; the stitched image is based on the target object image and similar target images In the splicing, the target object image is an image containing the target object, the similar target image is an image containing the similar target object, and the similarity between the similar target object and the target object is within a preset range.

Wherein, the target object may be a badge-like object, and the similar target object may be a target similar in shape or color to the badge, for example, a watch, a circular icon, and the like. Optionally, the target object images and similar target object images may be obtained from a preset database or a commonly used search engine.

In some embodiments, for a target detection model that recognizes badge targets, when acquiring training data, it is difficult to acquire a large amount of relevant training data for uncommon badge images, and a more accurate model can be obtained by image splicing. Multiple badge images of this type are used as training data.

It can be understood that, in addition to stitched images, the training sample data of the target detection model may also include a large number of multiple types of images such as target object images and similar target object images.

In the embodiment of the present disclosure, the number and richness of training samples can be increased by using the spliced images obtained by splicing target object images and similar target object images as training samples, and model training can be carried out through rich training sample images, which can make the model learn Get more information to improve the detection accuracy of the trained target detection model.

In one embodiment, the spliced image is obtained in the following manner:

Processing the image of the target object according to a preset method to obtain a processed image; the preset method includes at least one of sharpening, adding noise, filtering, color dithering, random filling and perspective transformation;

The processed image is used as the foreground image, and the similar target image is used as the background image for splicing.

It should be noted that the preset manner may also be any other image processing manner, which is not limited in this application.

In the embodiment of the present disclosure, the target object is processed in different ways to obtain images processed by each processing method, and then image stitching is performed, and the stitched images are used as training samples, which can increase the number and richness of training samples.

In one embodiment, the target detection method also includes:

During the training process of the initial target detection model, multiple test sample images are obtained, and the test sample images are input into the model to obtain output results corresponding to the test sample images;

The image corresponding to the detection error result in the output results corresponding to the target object image and the test sample image is spliced to obtain a new spliced image, and the detection error result is the result of detecting an image that does not contain the target object as containing the target object;

Obtain the annotation label of the new stitched image;

The first training sample data is updated with new stitched images and annotation labels.

Wherein, during the training process of the initial target detection model, a test sample image may also be obtained from a preset image database and input into the model for testing. The image corresponding to the detection error result output by the model is used as the background image, and is stitched with the target object image to obtain a new stitched image, which is added to the training sample database of the model. The model can be a model after adjusting the parameters of the initial target detection model through continuous iterations during the training process. The model at this time has not yet met the training end condition, and is not the target detection model that is finally trained.

In some embodiments, the tagged label of the new spliced image may be a manually tagged tag, or a tag obtained in other ways, which is not limited in this application.

In the embodiment of the present disclosure, the training sample data of the model can be increased by splicing the image corresponding to the detection error result corresponding to the test sample image in the model iteration process and the target object image, so that the training sample data can be realized. expansion.

In one embodiment, the target detection result is verified based on the position frame information, and the target verification result corresponding to the image to be detected is obtained, including:

Based on the position frame information, an image corresponding to the position frame is acquired in the image to be detected;

The image corresponding to the location frame is input into the pre-trained first classification model to obtain a classification result, and the target verification result corresponding to the image to be detected includes the classification result.

Wherein, the first classification model may include but not limited to classification models such as deep convolutional classification models, ResNet models, Inception models, and VGG models. Optionally, the classification result may be whether the target object is included or not, or may be the probability of including the target object or not including the target object, which is not limited in this application. Optionally, the classification result can be used as the final target verification result, so as to determine whether the image to be detected is an approved image.

Wherein, the image input into the first classification model may be an image framed by a position box in the detection result corresponding to the image to be detected.

In the embodiment of the present disclosure, the purpose of verification is achieved by classifying the location frame images corresponding to the target detection results, so as to improve the accuracy rate and recall rate of target detection.

In one embodiment, the first classification model is obtained by using multiple sets of second training samples to train the initial first classification model, and the second training samples include second sample images and second sample images corresponding to each second sample image. label; the second sample image includes the image corresponding to the position frame information of the target object in the first sample image, and the position frame image corresponding to the detection error result output by the initial target detection model during the training process; the second sample label uses To represent whether the target object is included in the second sample image.

In the embodiment of the present disclosure, the image corresponding to the position frame information of the target object in the first sample image is used as the positive sample of the first classification model, and the position corresponding to the detection error output by the initial target detection model during the training process is The frame image is used as a negative sample to train the first classification model, which can make the classification accuracy of the trained first classification model higher.

In one embodiment, the target detection result is verified based on the position frame information, and the target verification result corresponding to the image to be detected is obtained, including:

Based on the position frame information, an image corresponding to the position frame is acquired in the image to be detected;

Obtain the feature vector of the image corresponding to the location box;

Searching in a preset retrieval database based on the feature vector to obtain a retrieval result; the retrieval database includes the feature vector of the image of the target object;

According to the retrieval result, the target verification result corresponding to the image to be detected is determined.

In this embodiment, the feature vector of the image corresponding to the position frame can be extracted by the feature extraction module of the first classification model, and the feature vector can be used to search in the retrieval database. The retrieval method can include but not limited to approximate nearest neighbor search (Approximate Nearest Neighbor, ANN). The retrieval database is a pre-established database containing feature vectors of images of various types of target objects. If the similarity between the retrieved position frame image and the image in the retrieved database is greater than a preset threshold, it is considered that the image to be processed contains the target object.

In the embodiment of the present disclosure, the accuracy of target detection can be improved by checking whether the target object is included in the image to be detected by searching in the search database. The data in the retrieval database in the embodiments of the present disclosure can be continuously updated, and feature vectors of newly added target object images can be continuously added to the retrieval database, thereby improving the recall ability of newly added target object images.

In one embodiment, the target detection result is verified based on the position frame information, and the target verification result corresponding to the image to be detected is obtained, including:

Based on the position frame information, an image corresponding to the position frame is acquired in the image to be detected;

Obtain feature information of the image corresponding to the position frame, and determine a target verification result corresponding to the image to be detected based on the feature information;

Feature information includes at least one of the following:

The shape feature of the object contained in the image, the color feature of the object contained in the image and the text information in the image.

In the embodiment of the present disclosure, the visual features of the position frame image can also be extracted through the computer vision algorithm, including but not limited to the shape features of the objects contained in the image, the color features of the objects contained in the image, and the text information in the image. At least one of these features is used to determine whether the target object is contained in the location frame image, thereby improving the accuracy of the target detection result.

In the technical solution of the present disclosure, an audit exemption mechanism may also be included. For the image of the target object included in the target verification result, it is further determined whether to perform audit exemption. For details, see the following embodiments.

In one embodiment, the target object includes a badge object, and the target detection method further includes:

When the target verification result is that the image to be detected contains a badge-like object, input the image to be detected into the preset second classification model, and determine that the image to be detected is a book cover image according to the output result of the second classification model , and determine the image to be detected as the approved image.

In the embodiment of the present disclosure, when the target object is a badge object, the second classification model may be a model for classifying book cover images and certificate images. Input the image to be detected or the position frame image of the image to be detected into the second classification model to determine whether it is a book cover image. The book cover image can be an image that is exempt from review. The inspection image is approved.

In the embodiment of the present disclosure, the output result of the second classification model is used to determine whether the image to be detected can be exempted from examination, which can meet the examination exemption requirement for the book cover image.

In one embodiment, the target detection method also includes:

When it is determined that the image to be detected is not a book cover image according to the output result of the second classification model, the text information in the image to be detected is obtained, and when the image to be detected is determined to be a legal document image according to the text information, the image to be detected is Determined as approved image.

In the embodiment of the present disclosure, when it is determined according to the output result of the second classification model that the image to be detected is not a book cover image, according to the text information in the image to be detected, through optical character recognition (Optical Character Recognition, OCR) and other methods, the The text information in the image to be detected is identified to determine whether it is a legal certificate name, thereby determining whether the image to be detected is a legal certificate image, and if it is a legal certificate image, the image to be detected is determined to be an approved image.

In the embodiment of the present disclosure, it is determined whether the image to be detected can be exempted from examination by means of character recognition, which can meet the examination exemption requirements for legal document images.

In one embodiment, the target detection method also includes:

When it is determined according to the text information that the image to be detected is not a legal document image, obtain the user qualification information corresponding to the image to be detected, determine whether the user corresponding to the image to be detected has audit exemption authority according to the user qualification, and if so, determine the image to be detected Approved image for review.

In the embodiment of the present disclosure, for a specific user, an audit exemption authority may be granted, and an image uploaded by the user may be exempted from audit. The user qualification information can be bound to the image uploaded by the user, and the user qualification information of the user corresponding to the image can be queried through the identification information of the image, etc. According to the user qualification information, it can be determined whether the user corresponding to the image to be detected has the audit exemption authority.

In the embodiment of the present disclosure, by verifying the user qualification information, it is determined whether the image to be detected can be exempted from auditing, which can meet the auditing exemption requirements for images uploaded by users with auditing exemption authority.

In one embodiment, the target detection method also includes:

When it is determined according to the user qualification that the user corresponding to the image to be detected does not have audit exemption authority, the image to be detected is determined to be an image that has failed the review.

In the embodiment of the present disclosure, for an image containing a badge-type target object sent by a user who does not have the audit exemption authority, it is determined to be an image that fails the audit.

In the target detection method in the disclosed technical solution, in the target detection results output by the target detection model, the target detection results containing position frame information and whose confidence level is less than the preset confidence level threshold are verified, so as to improve the accuracy of image review. Precision and recall for images containing the target object.

FIG. 2 is a schematic diagram of an acquisition process of training samples of a target detection model in an embodiment of the present disclosure. The target object in this embodiment is a badge object. As shown in Figure 2, the image containing the badge is obtained as a positive sample image, and the sample is labeled ("positive sample capture and labeling" as shown in the figure). Obtain an image similar to the badge image as a similar target image (as shown in the figure "Badge shape similar image capture"); search for images similar to the badge image in the preset database (as shown in the figure "Advertising material library search"), use the badge image and images similar to the badge image to stitch together to obtain a stitched image (as shown in the "sample generation and amplification" shown in the figure), and use the badge image, the image similar to the badge, and the stitched image as the target The first training sample data of the detection model ("initial training data" shown in the figure).

Fig. 3 is a schematic diagram of an iterative process of a target detection model in an embodiment of the present disclosure. As shown in Figure 3, the target object image and similar target object images are obtained, and stitched, and the sample label of the stitched image (as shown in the figure "initial annotation, generated data") is obtained, and the stitched image and sample label are used as The training data set of the target detection model, and iterative training of the target detection model, in the iterative process of the model, the test sample image is input into the model, and the position frame image corresponding to the detection error result output by the model is used as the background image (as shown in Fig. The "negative example collection" shown in the figure) is spliced with the target object image to obtain a new spliced image, which is added to the training sample set of the target detection model ("sample generation and amplification" shown in the figure), using continuous The updated training data set trains the model ("model iteration" as shown in the figure), until the preset training end condition is met, and the target detection model is obtained ("model release" as shown in the figure).

FIG. 4 is a schematic diagram of an image review system in an embodiment of the present disclosure. As shown in FIG. 4 , in this embodiment, the target object image is a badge image. The image review system includes: a target detection unit, a secondary verification unit, and a special scene recognition and exemption unit. The image to be detected is input to the target detection unit, and the badge detection is performed by a sensitive badge target detector (which can be the target detection model in the disclosed technical solution), and the badge detection result is obtained. If the probability of including the badge is less than the preset threshold (as shown in the figure Shown as "less than the threshold"), the image to be detected does not contain a badge, and the image to be processed is determined to be an approved image. If the probability that the badge detection result contains a badge is greater than or equal to the preset threshold, it means that the image to be detected may contain a badge, and the target detection result contains location frame information, and the confidence corresponding to the target detection result is less than the preset confidence threshold The images to be processed are verified, and the position frame images are extracted from these images, and input into the secondary verification unit for verification. The specific verification methods can include three types: the first search library search, and the specific implementation process includes: extracting the position Frame the feature vector of the image, and perform compression processing, use the processed feature vector to search in the preset search database, get the search result, and determine whether to include the badge according to the search result. The second type of morphological feature detection, the specific implementation process includes: extracting feature information from the position frame image, and determining the image according to the shape features of the object contained in the image, the color features of the object contained in the image, and the text information in the image. Whether to include a badge in . The third method is to determine whether the image contains a badge based on the classification result of the badge recognition classifier (the first classification model in the technical solution of the present disclosure). If it is determined according to the verification result of the secondary verification unit that the probability of including the badge in the image is less than the preset threshold ("less than the threshold" as shown in the figure), the approval is passed directly. When the target verification result is that the image to be detected contains a badge image, the image to be detected is input into the special scene recognition and exemption unit, and the certificate book classifier (the second classification model in the disclosed technical solution) is used for classification and recognition. If it is a book cover image, it is approved. If it is not a book cover image, the text information in the image to be detected is obtained, and the name of the legal certificate is verified. If it is a legal certificate image, the review is passed. If it is not a legal document image, then obtain the user qualification information corresponding to the image to be detected, and perform qualification authorization verification. If the user who uploaded the image has the qualification for exemption from review, the review is passed; otherwise, the image to be detected is determined to be an image that has not passed the review.

FIG. 5 is a schematic diagram of a method for training a target detection model in an embodiment of the present disclosure. As shown in Figure 5, the training method of the target detection model may include:

Step S501, acquiring multiple sets of first training sample data;

Wherein, the first training sample data includes the first sample image and the first sample label corresponding to the first sample image, and the first sample label is used to represent whether the first sample image contains the target object and the position of the target object frame information, the first sample image includes a mosaic image; the mosaic image is obtained by stitching based on the target object image and similar target images, the target object image is an image containing the target object, the similar target image is an image containing similar target objects, and the similar The similarity between the target object and the target object is within a preset range.

In step S502, the initial target detection model is trained based on multiple sets of first training sample data until the preset training end condition is satisfied, and the target detection model of any embodiment of the present disclosure is obtained.

The training method of the target detection model in the disclosed technical solution can increase the number and richness of training samples by using the spliced images obtained by splicing target object images and similar target object images as training samples. Model training can enable the model to learn more information to improve the detection accuracy of the trained target detection model.

In one embodiment, the spliced image is obtained in the following manner:

Processing the image of the target object according to a preset method to obtain a processed image; the preset method includes at least one of sharpening, adding noise, filtering, color dithering, random filling and perspective transformation;

The processed image is used as the foreground image, and the similar target image is used as the background image for splicing.

It should be noted that the preset manner may also be any other image processing manner, which is not limited in this application.

In the embodiment of the present disclosure, the target object is processed in different ways to obtain images processed by each processing method, and then image stitching is performed, and the stitched images are used as training samples, which can increase the number and richness of training samples.

In one embodiment, the training method of target detection model also includes:

During the training process of the initial target detection model, multiple test sample images are obtained, and the test sample images are input into the model to obtain output results corresponding to the test sample images;

The image corresponding to the detection error result in the output results corresponding to the target object image and the test sample image is spliced to obtain a new spliced image, and the detection error result is the result of detecting an image that does not contain the target object as containing the target object;

Obtain the annotation label of the new stitched image;

The first training sample data is updated with new stitched images and annotation labels.

Wherein, during the training process of the initial target detection model, a test sample image may also be obtained from a preset image database and input into the model for testing. The image corresponding to the detection error result output by the model is used as the background image, and is stitched with the target object image to obtain a new stitched image, which is added to the training sample database of the model. The model can be a model after adjusting the parameters of the initial target detection model through continuous iterations during the training process. The model at this time has not yet met the training end condition, and is not the target detection model that is finally trained.

In some embodiments, the tagged label of the new spliced image may be a manually tagged tag, or a tag obtained in other ways, which is not limited in this application.

In the embodiment of the present disclosure, the training sample data of the model can be increased by splicing the image corresponding to the detection error result corresponding to the test sample image in the model iteration process and the target object image, so that the training sample data can be realized. expansion.

FIG. 6 is a schematic diagram of an object detection device in an embodiment of the present disclosure. As shown in Figure 6, the target detection device may include:

A detection image acquisition module 601, configured to obtain an image to be detected;

A target detection module 602, configured to input the image to be detected into a pre-trained target detection model to obtain a target detection result corresponding to the image to be detected;

The position frame obtaining module 603 is used to obtain the position frame information in the target detection result when the target detection result satisfies the preset condition; wherein, the preset condition is that the target detection result contains the position frame information, and the target detection result corresponds to The confidence of is less than the preset confidence threshold;

The result verification module 604 is configured to verify the target detection result based on the location frame information, and obtain a target verification result corresponding to the image to be detected.

In one embodiment, the target detection model is obtained by using multiple sets of first training sample data to train the initial target detection model. The first training sample data includes the first sample image and the first sample image corresponding to the first sample image. A sample label, the first sample label is used to represent whether the first sample image contains the target object and the location frame information of the target object;

The first sample image includes a mosaic image; the mosaic image is obtained based on the mosaic of the target object image and the similar target image, the target object image is an image containing the target object, the similar target image is an image containing the similar target object, and the similar target object and The similarity of the target object is within a preset range.

In one embodiment, the spliced image is obtained in the following manner:

Processing the image of the target object according to a preset method to obtain a processed image; the preset method includes at least one of sharpening, adding noise, filtering, color dithering, random filling and perspective transformation;

The processed image is used as the foreground image, and the similar target image is used as the background image for splicing.

FIG. 7 is a schematic diagram of a sample update module in an embodiment of the present disclosure. As shown in Figure 7, in one embodiment, the target detection device also includes a sample update module, and the sample update module includes:

The test result acquisition unit 701 is used to obtain a plurality of test sample images during the training process of the initial target detection model, input the test sample images into the model, and obtain the corresponding output results of the test sample images;

The image stitching unit 702 is configured to stitch the image corresponding to the result of the detection error among the output results corresponding to the target object image and the test sample image to obtain a new stitched image, and the result of the detection error is to detect an image that does not contain the target object. for results containing the target object;

A label obtaining unit 703, configured to obtain an annotation label of a new spliced image;

An updating unit 704, configured to update the first training sample data with the new spliced image and label.

In one embodiment, the result verification module 604 is specifically used for:

Based on the position frame information, an image corresponding to the position frame is acquired in the image to be detected;

The image corresponding to the location frame is input into the pre-trained first classification model to obtain a classification result, and the target verification result corresponding to the image to be detected includes the classification result.

In one embodiment, the first classification model is obtained by using multiple sets of second training samples to train the initial first classification model, and the second training samples include second sample images and second sample images corresponding to each second sample image. label; the second sample image includes the image corresponding to the position frame information of the target object in the first sample image, and the position frame image corresponding to the detection error result output by the initial target detection model during the training process; the second sample label uses To represent whether the target object is included in the second sample image.

In one embodiment, the result verification module 604 is specifically used for:

Based on the position frame information, an image corresponding to the position frame is acquired in the image to be detected;

Obtain the feature vector of the image corresponding to the location frame;

Searching in a preset retrieval database based on the feature vector to obtain a retrieval result; the retrieval database includes the feature vector of the image of the target object;

According to the retrieval result, the target verification result corresponding to the image to be detected is determined.

In one embodiment, the result verification module 604 is specifically used for:

Based on the position frame information, an image corresponding to the position frame is acquired in the image to be detected;

Obtain feature information of the image corresponding to the position frame, and determine a target verification result corresponding to the image to be detected based on the feature information;

Feature information includes at least one of the following:

The shape feature of the object contained in the image, the color feature of the object contained in the image and the text information in the image.

In one embodiment, the target object includes a badge-like object, and the target detection device further includes an image classification module, configured to:

When the target verification result is that the image to be detected contains a badge-like object, input the image to be detected into the preset second classification model, and determine that the image to be detected is a book cover image according to the output result of the second classification model , and determine the image to be detected as the approved image.

In one embodiment, the target detection device also includes a character recognition module, configured to:

When it is determined that the image to be detected is not a book cover image according to the output result of the second classification model, the text information in the image to be detected is obtained, and when the image to be detected is determined to be a legal document image according to the text information, the image to be detected is Determined as approved image.

In one embodiment, the target detection device also includes a qualification verification module, which is used for:

When it is determined according to the text information that the image to be detected is not a legal document image, obtain the user qualification information corresponding to the image to be detected, determine whether the user corresponding to the image to be detected has audit exemption authority according to the user qualification, and if so, determine the image to be detected Approved image for review.

In one embodiment, the target detection device also includes a result determination module, configured to:

When it is determined according to the user qualification that the user corresponding to the image to be detected does not have audit exemption authority, the image to be detected is determined to be an image that has failed the review.

In the target detection device of the embodiment of the present disclosure, in the target detection results output by the target detection model, the target detection results containing the location frame information and whose confidence is less than the preset confidence threshold are verified, so as to improve the accuracy of image review. rate and recall of images containing the target object.

FIG. 8 is a schematic diagram of a training device for a target detection model in an embodiment of the present disclosure. As shown in Figure 8, the training device for target detection can include:

A sample acquisition module 801, configured to acquire multiple sets of first training sample data;

The model training module 802 is configured to train the initial target detection model based on multiple sets of first training sample data until the preset training end condition is met, and obtain the target detection model of any embodiment of the present disclosure;

Wherein, the first training sample data includes the first sample image and the first sample label corresponding to the first sample image, and the first sample label is used to represent whether the first sample image contains the target object and the position of the target object frame information, the first sample image includes a mosaic image; the mosaic image is obtained by stitching based on the target object image and similar target images, the target object image is an image containing the target object, the similar target image is an image containing similar target objects, and the similar The similarity between the target object and the target object is within a preset range.

In one embodiment, the spliced image is obtained in the following manner:

Processing the image of the target object according to a preset method to obtain a processed image; the preset method includes at least one of sharpening, adding noise, filtering, color dithering, random filling and perspective transformation;

The processed image is used as the foreground image, and the similar target image is used as the background image for splicing.

In one embodiment, the training device of the target detection model also includes a sample update module, and the sample update module includes:

The test result acquisition unit is used to obtain a plurality of test sample images during the training process of the initial target detection model, input the test sample images into the model, and obtain the corresponding output results of the test sample images;

The image splicing unit is used to splice the image corresponding to the result of the detection error in the output results corresponding to the target object image and the test sample image to obtain a new spliced image, and the result of the detection error is to detect the image that does not contain the target object as Contains results for the target object;

A label acquisition unit, configured to acquire an annotation label of a new spliced image;

An update unit, configured to update the first training sample data with new stitched images and annotation labels.

The training device of the target detection model in the disclosed technical solution can increase the number and richness of training samples by using the spliced images obtained by splicing target object images and similar target object images as training samples. Model training can enable the model to learn more information to improve the detection accuracy of the trained target detection model.

For functions of each unit, module, or submodule in each device in the embodiments of the present disclosure, reference may be made to the corresponding descriptions in the foregoing method embodiments, and details are not repeated here.

According to the embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

FIG. 9 shows a schematic block diagram of an example electronic device 900 that may be used to implement embodiments of the present disclosure. Electronic device is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are by way of example only, and are not intended to limit implementations of the disclosure described and/or claimed herein.

As shown in FIG. 9 , an electronic device 900 includes a computing unit 901 that can be executed according to a computer program stored in a read-only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903 Various appropriate actions and treatments. In the RAM 903, various programs and data necessary for the operation of the electronic device 900 can also be stored. The computing unit 901 , ROM 902 , and RAM 903 are connected to each other through a bus 904 . An input-output (I/O) interface 905 is also connected to the bus 904 .

Multiple components in the electronic device 900 are connected to the I/O interface 905, including: an input unit 906, such as a keyboard, a mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a magnetic disk, an optical disk etc.; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 909 allows the electronic device 900 to exchange information/data with other devices through a computer network such as the Internet and/or various telecommunication networks.

The computing unit 901 may be various general-purpose and/or special-purpose processing components having processing and computing capabilities. Some examples of computing units 901 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 executes various methods and processes described above, such as object detection methods. For example, in some embodiments, an object detection method or a method for training an object detection model may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 908 . In some embodiments, part or all of the computer program may be loaded and/or installed on the electronic device 900 via the ROM 902 and/or the communication unit 909 . When the computer program is loaded into RAM 903 and executed by computing unit 901, one or more steps of the object detection method described above may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured in any other appropriate way (for example, by means of firmware) to execute the object detection method.

Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips Implemented in a system of systems (SOC), load programmable logic device (CPLD), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs executable and/or interpreted on a programmable system including at least one programmable processor, the programmable processor Can be special-purpose or general-purpose programmable processor, can receive data and instruction from storage system, at least one input device, and at least one output device, and transmit data and instruction to this storage system, this at least one input device, and this at least one output device an output device.

Program codes for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special purpose computer, or other programmable data processing devices, so that the program codes, when executed by the processor or controller, make the functions/functions specified in the flow diagrams and/or block diagrams Action is implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media would include one or more wire-based electrical connections, portable computer discs, hard drives, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.

To provide for interaction with the 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 (eg, a mouse or a trackball) through which a user can provide input to the computer. Other kinds 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 can be in any form (including Input from the user may be received through acoustic input, voice input, or tactile input.

The systems and techniques described herein can 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., as a a user computer having a graphical user interface or web browser through which a user can interact with embodiments of the systems and techniques described herein), or including such backend components, middleware components, Or any combination of front-end components in a computing system. The components of the system can be interconnected by any form or medium of digital data communication, eg, a communication network. Examples of communication networks include: Local Area Network (LAN), Wide Area Network (WAN) and the Internet.

A computer system may include clients and servers. Clients and servers are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by computer programs running on the respective computers and having a client-server relationship to each other.

It should be understood that steps may be reordered, added or deleted using the various forms of flow shown above. For example, each step described in the present disclosure may be executed in parallel, sequentially, or in a different order, as long as the desired result of the technical solution disclosed in the present disclosure can be achieved, no limitation is imposed herein.

The specific implementation manners described above do not limit the protection scope of the present disclosure. It should be apparent to those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made depending on design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present disclosure shall be included within the protection scope of the present disclosure.

Claims (26)

1. A method of target detection, the method comprising:

acquiring an image to be detected;

inputting the image to be detected into a pre-trained target detection model to obtain a target detection result corresponding to the image to be detected;

acquiring position frame information in the target detection result under the condition that the target detection result meets a preset condition; the preset condition is that the target detection result contains position frame information, and the confidence coefficient corresponding to the target detection result is smaller than a preset confidence coefficient threshold value;

verifying the target detection result based on the position frame information to obtain a target verification result corresponding to the image to be detected;

the verifying the target detection result based on the position frame information to obtain a target verification result corresponding to the image to be detected comprises the following steps:

Acquiring an image corresponding to the position frame from the image to be detected based on the position frame information;

inputting the image corresponding to the position frame into a pre-trained first classification model to obtain a classification result, wherein a target verification result corresponding to the image to be detected comprises the classification result, and the classification result comprises a target object or a target object not;

or,

acquiring a feature vector of an image corresponding to the position frame;

searching in a preset search database based on the feature vector to obtain a search result; the retrieval database comprises feature vectors of the target object image;

determining a target verification result corresponding to the image to be detected according to the retrieval result;

if the similarity between the image corresponding to the position frame and the image in the search database is larger than a preset threshold value, the image to be detected contains a target object;

or,

acquiring characteristic information of an image corresponding to the position frame, and determining a target verification result corresponding to the image to be detected based on the characteristic information;

the determining the target verification result corresponding to the image to be detected based on the characteristic information comprises the following steps: determining whether the position frame image contains a target object or not according to the characteristic information;

The characteristic information includes at least one of: the shape characteristics of the objects contained in the image, the color characteristics of the objects contained in the image, and the text information in the image.

2. The method according to claim 1, wherein the target detection model is obtained by training an initial target detection model with a plurality of sets of first training sample data, the first training sample data including a first sample image and a first sample tag corresponding to the first sample image, the first sample tag being used for representing whether the first sample image includes a target object and position frame information of the target object;

the first sample image comprises a stitched image; the stitching image is obtained by stitching based on a target object image and a similar target image, the target object image is an image containing a target object, the similar target image is an image containing a similar target object, and the similarity between the similar target object and the target object is within a preset range.

3. The method of claim 2, wherein the stitched image is obtained by:

processing the target object image according to a preset mode to obtain a processed image; the preset mode comprises at least one of sharpening, noise adding, filtering, color dithering, random filling and perspective transformation;

And taking the processed image as a foreground image, taking the similar target image as a background image, and performing splicing processing.

4. A method according to claim 2 or 3, characterized in that the method further comprises:

in the training process of the initial target detection model, a plurality of test sample images are obtained, the test sample images are input into the model, and output results corresponding to the test sample images are obtained;

splicing the images corresponding to the error detection results in the output results corresponding to the target object images and the test sample images to obtain new spliced images, wherein the error detection results are the results of detecting the images which do not contain the target object as the results containing the target object;

acquiring a labeling label of the new spliced image;

and updating the first training sample data by using the new spliced image and the labeling label.

5. The method of claim 4, wherein the first classification model is obtained by training an initial first classification model with a plurality of sets of second training samples, the second training samples comprising second sample images and second sample labels corresponding to each second sample image; the second sample image comprises an image corresponding to the position frame information of the target object in the first sample image and a position frame image corresponding to the result of the detection error output by the initial target detection model in the training process; the second sample tag is used for representing whether the second sample image contains a target object.

6. The method of claim 1, wherein the target object comprises a badge class object, the method further comprising:

and inputting the image to be detected into a preset second classification model under the condition that the target verification result is that the image to be detected contains badge-like objects, and determining the image to be detected as an audit passing image under the condition that the image to be detected is determined to be a book cover image according to the output result of the second classification model.

7. The method of claim 6, wherein the method further comprises:

and under the condition that the image to be detected is not the book cover image according to the output result of the second classification model, acquiring text information in the image to be detected, and under the condition that the image to be detected is a legal document image according to the text information, determining the image to be detected as a checking passing image.

8. The method of claim 7, wherein the method further comprises:

under the condition that the image to be detected is not a legal document image according to the text information, obtaining user qualification information corresponding to the image to be detected, determining whether a user corresponding to the image to be detected has checking and exempting permission according to the user qualification, and if so, determining the image to be detected as a checking and passing image.

9. The method of claim 8, wherein the method further comprises:

and determining the image to be detected as an audit failed image when the user corresponding to the image to be detected does not have audit exemption authority according to the user qualification.

10. A method of training a target detection model, the method comprising:

acquiring a plurality of groups of first training sample data, and training an initial target detection model based on the plurality of groups of first training sample data until a preset training ending condition is met to obtain the target detection model according to any one of claims 1-9;

the first training sample data comprises a first sample image and a first sample label corresponding to the first sample image, wherein the first sample label is used for representing whether the first sample image contains a target object and position frame information of the target object or not, and the first sample image comprises a spliced image; the stitching image is obtained by stitching based on a target object image and a similar target image, the target object image is an image containing a target object, the similar target image is an image containing a similar target object, and the similarity between the similar target object and the target object is within a preset range.

11. The method of claim 10, wherein the stitched image is obtained by:

processing the target object image according to a preset mode to obtain a processed image; the preset mode comprises at least one of sharpening, noise adding, filtering, color dithering, random filling and perspective transformation;

and taking the processed image as a foreground image, taking the similar target image as a background image, and performing splicing processing.

12. The method according to any one of claims 10 or 11, further comprising:

in the training process of the initial target detection model, a plurality of test sample images are obtained, the test sample images are input into the model, and output results corresponding to the test sample images are obtained;

splicing the images corresponding to the error detection results in the output results corresponding to the target object images and the test sample images to obtain new spliced images, wherein the error detection results are the results of detecting the images which do not contain the target object as the results containing the target object;

acquiring a labeling label of the new spliced image;

and updating the first training sample data by using the new spliced image and the labeling label.

13. An object detection apparatus, the apparatus comprising:

the detection image acquisition module is used for acquiring an image to be detected;

the target detection module is used for inputting the image to be detected into a pre-trained target detection model to obtain a target detection result corresponding to the image to be detected;

the position frame acquisition module is used for acquiring position frame information in the target detection result under the condition that the target detection result meets the preset condition; the preset condition is that the target detection result contains position frame information, and the confidence coefficient corresponding to the target detection result is smaller than a preset confidence coefficient threshold value;

the result checking module is used for checking the target detection result based on the position frame information to obtain a target checking result corresponding to the image to be detected;

the result checking module is specifically configured to obtain an image corresponding to a position frame from the image to be detected based on the position frame information; inputting the image corresponding to the position frame into a pre-trained first classification model to obtain a classification result, wherein a target verification result corresponding to the image to be detected comprises the classification result, and the classification result comprises a target object or a target object not; or, obtaining the feature vector of the image corresponding to the position frame; searching in a preset search database based on the feature vector to obtain a search result; the retrieval database comprises feature vectors of the target object image; determining a target verification result corresponding to the image to be detected according to the retrieval result; if the similarity between the image corresponding to the position frame and the image in the search database is larger than a preset threshold value, the image to be detected contains a target object; or, acquiring characteristic information of an image corresponding to the position frame, and determining a target verification result corresponding to the image to be detected based on the characteristic information; the determining the target verification result corresponding to the image to be detected based on the characteristic information comprises the following steps: determining whether the position frame image contains a target object or not according to the characteristic information; the characteristic information includes at least one of: the shape characteristics of the objects contained in the image, the color characteristics of the objects contained in the image, and the text information in the image.

14. The apparatus of claim 13, wherein the target detection model is obtained by training an initial target detection model with a plurality of sets of first training sample data, the first training sample data including a first sample image and a first sample tag corresponding to the first sample image, the first sample tag being used to represent whether the first sample image includes a target object and position frame information of the target object;

the first sample image comprises a stitched image; the stitching image is obtained by stitching based on a target object image and a similar target image, the target object image is an image containing a target object, the similar target image is an image containing a similar target object, and the similarity between the similar target object and the target object is within a preset range.

15. The apparatus of claim 14, wherein the stitched image is obtained by:

processing the target object image according to a preset mode to obtain a processed image; the preset mode comprises at least one of sharpening, noise adding, filtering, color dithering, random filling and perspective transformation;

And taking the processed image as a foreground image, taking the similar target image as a background image, and performing splicing processing.

16. The apparatus of claim 14 or 15, further comprising a sample update module, the sample update module comprising:

the test result acquisition unit is used for acquiring a plurality of test sample images in the training process of the initial target detection model, inputting the test sample images into the model, and obtaining an output result corresponding to the test sample images;

the image stitching unit is used for stitching the images corresponding to the detection error results in the output results corresponding to the target object images and the test sample images to obtain new stitched images, wherein the detection error results are the results of detecting the images which do not contain the target object as the results containing the target object;

the label acquisition unit is used for acquiring the labeling label of the new spliced image;

and the updating unit is used for updating the first training sample data by utilizing the new spliced image and the labeling label.

17. The apparatus of claim 16, wherein the first classification model is trained on an initial first classification model using a plurality of sets of second training samples, the second training samples comprising second sample images and second sample labels corresponding to each second sample image; the second sample image comprises an image corresponding to the position frame information of the target object in the first sample image and a position frame image corresponding to the result of the detection error output by the initial target detection model in the training process; the second sample tag is used for representing whether the second sample image contains a target object.

18. The apparatus of claim 17, wherein the target object comprises a badge class object, the apparatus further comprising an image classification module to:

and inputting the image to be detected into a preset second classification model under the condition that the target verification result is that the image to be detected contains badge-like objects, and determining the image to be detected as an audit passing image under the condition that the image to be detected is determined to be a book cover image according to the output result of the second classification model.

19. The apparatus of claim 18, further comprising a text recognition module for:

and under the condition that the image to be detected is not the book cover image according to the output result of the second classification model, acquiring text information in the image to be detected, and under the condition that the image to be detected is a legal document image according to the text information, determining the image to be detected as a checking passing image.

20. The apparatus of claim 19, further comprising a qualification audit module for:

under the condition that the image to be detected is not a legal document image according to the text information, obtaining user qualification information corresponding to the image to be detected, determining whether a user corresponding to the image to be detected has checking and exempting permission according to the user qualification, and if so, determining the image to be detected as a checking and passing image.

21. The apparatus of claim 20, further comprising a result determination module configured to:

and determining the image to be detected as an audit failed image when the user corresponding to the image to be detected does not have audit exemption authority according to the user qualification.

22. A training apparatus for a target detection model, the apparatus comprising:

the sample acquisition module is used for acquiring a plurality of groups of first training sample data;

the model training module is used for training the initial target detection model based on the multiple groups of first training sample data until a preset training ending condition is met, so as to obtain the target detection model according to any one of claims 1-9;

the first training sample data comprises a first sample image and a first sample label corresponding to the first sample image, wherein the first sample label is used for representing whether the first sample image contains a target object and position frame information of the target object or not, and the first sample image comprises a spliced image; the stitching image is obtained by stitching based on a target object image and a similar target image, the target object image is an image containing a target object, the similar target image is an image containing a similar target object, and the similarity between the similar target object and the target object is within a preset range.

23. The apparatus of claim 22, wherein the stitched image is obtained by:

processing the target object image according to a preset mode to obtain a processed image; the preset mode comprises at least one of sharpening, noise adding, filtering, color dithering, random filling and perspective transformation;

and taking the processed image as a foreground image, taking the similar target image as a background image, and performing splicing processing.

24. The apparatus of any one of claims 22 or 23, further comprising a sample update module comprising:

the test result acquisition unit is used for acquiring a plurality of test sample images in the training process of the initial target detection model, inputting the test sample images into the model, and obtaining an output result corresponding to the test sample images;

the image stitching unit is used for stitching the images corresponding to the detection error results in the output results corresponding to the target object images and the test sample images to obtain new stitched images, wherein the detection error results are the results of detecting the images which do not contain the target object as the results containing the target object;

The label acquisition unit is used for acquiring the labeling label of the new spliced image;

and the updating unit is used for updating the first training sample data by utilizing the new spliced image and the labeling label.

25. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-12.

26. A non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1-12.