WO2023284307A1 - Image processing method and apparatus, and electronic device, storage medium and computer program product - Google Patents

Abstract

Provided in the embodiments of the present disclosure are an image processing method and apparatus, and an electronic device, a storage medium and a computer program product. The method comprises: in response to a first object recognition instruction, acquiring a first object image of a target object that is collected by a local device, and a target reference image, wherein the target reference image is obtained by updating a registration image on the basis of a local object image, the registration image is an image of the target object that is collected by a remote device, and the registration image is used for performing object registration; and performing object recognition on the first object image on the basis of the target reference image, so as to obtain a first object recognition result.

Description

Image processing method and device, electronic device, storage medium and computer program product

Cross References to Related Applications

The embodiment of the present disclosure is based on the Chinese patent application with the application number 202110805912.6, the application date is July 16, 2021, and the application name is "image processing method, device, electronic equipment and storage medium", and requires the priority of the Chinese patent application Right, the entire content of this Chinese patent application is hereby incorporated into this disclosure as a reference.

technical field

The present disclosure relates to but not limited to the technical field of computer vision, and in particular relates to image processing methods and devices, electronic equipment, storage media and computer program products.

Background technique

With the development of computer vision technology, various applications based on image recognition have also been widely used, such as access control based on face recognition, attendance management, cashier payment, etc. In related technologies, due to differences among different devices, the accuracy rate, pass rate and security of object recognition are not high.

Contents of the invention

Embodiments of the present disclosure provide an image processing method and device, electronic equipment, a storage medium, and a computer program product.

An embodiment of the present disclosure provides an image processing method, including:

In response to the first object recognition instruction, acquire the first object image and the target reference image of the target object collected by the local device, the target reference image is obtained by updating the registered image based on the local object image, and the registered image is collected by the remote device an image of the target object, the registration image is used for object registration;

Perform object recognition on the first object image based on the target reference image to obtain a first object recognition result.

An embodiment of the present disclosure provides an image processing device, including:

The first image acquisition part is configured to execute in response to the first object recognition instruction, acquire the first object image and the target reference image of the target object captured by the local device, and the target reference image is obtained by updating the registration image based on the local object image Wherein, the registration image is an image of the target object collected by a remote device, and the registration image is used for object registration;

The first object recognition part is configured to perform object recognition on the first object image based on the target reference image to obtain a first object recognition result.

An embodiment of the present disclosure provides an electronic device, including: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement the above method.

An embodiment of the present disclosure provides a computer-readable storage medium, and when instructions in the storage medium are executed by a processor of the electronic device, the electronic device can execute the above method.

An embodiment of the present disclosure provides a computer program product, where the computer program product includes a computer program or an instruction, and when the computer program or instruction is run on an electronic device, the electronic device is made to execute the above method.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure.

Description of drawings

The accompanying drawings here are incorporated into the specification and constitute a part of the specification, show embodiments consistent with the disclosure, and are used together with the description to explain the principle of the disclosure, and do not constitute an improper limitation of the disclosure.

FIG. 1 is a schematic flowchart of an image processing method provided by an embodiment of the present disclosure;

FIG. 2 is a schematic flow diagram of obtaining a target reference image in advance according to an embodiment of the present disclosure;

FIG. 3 is a schematic flowchart of performing object recognition on a second object image based on a registered image to obtain a second object recognition result provided by an embodiment of the present disclosure;

FIG. 4 is a schematic flow chart of updating a target reference image provided by an embodiment of the present disclosure;

FIG. 5 is a schematic flowchart of obtaining a target reference image in advance according to an embodiment of the present disclosure;

FIG. 6 is a schematic diagram of the composition and structure of an image processing device provided by an embodiment of the present disclosure;

FIG. 7 is a schematic diagram of a hardware structure of an electronic device provided by an embodiment of the present disclosure.

detailed description

In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.

It should be noted that the terms "first" and "second" in the specification and claims of the present disclosure and the above drawings are used to distinguish similar objects, but not necessarily used to describe a specific sequence or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances such that the embodiments of the disclosure described herein can be practiced in sequences other than those illustrated or described herein. The implementations described in the following exemplary examples do not represent all implementations consistent with the present disclosure. Rather, they are examples of apparatuses and methods consistent with aspects of the present disclosure as recited in the appended claims.

It should be understood that in this disclosure, "at least one (item)" means one or more, "multiple" means two or more, and "at least two (items)" means two or three And three or more, "and/or", is used to describe the association relationship of associated objects, indicating that there can be three types of relationships, for example, "A and/or B" can mean: only A exists, only B exists, and A exists at the same time and B, where A and B can be singular or plural. The character "/" can indicate that the contextual objects are an "or" relationship, which refers to any combination of these items, including any combination of single items (items) or plural items (items). For example, at least one item (piece) of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c ", where a, b, c can be single or multiple. The character "/" can also represent a division sign in mathematical operations, for example, a/b=a divided by b; 6/3=2. "At least one of the following" or similar expressions.

Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present disclosure. The appearances of this phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is understood explicitly and implicitly by those skilled in the art that the embodiments described herein can be combined with other embodiments.

With the development of computer vision technology, various applications based on image recognition have also been widely used, such as access control based on face recognition, attendance management, cashier payment, etc. At present, algorithms for face recognition and matching in related technologies are often trained based on images collected by face recognition equipment. However, in practical applications, it often happens that the registration image on the face recognition device and the image to be identified and matched come from different devices. For example, in hotels, factories, dormitories, campuses and other scenarios, the use of face recognition equipment is often: the user uses an image on the mobile phone to perform remote registration in advance as a registration map; then, when the face recognition device needs to be used, The image captured by the user standing in front of the face recognition device is used as the object image. In this scenario, due to the differences between different devices, the optical properties of images such as brightness and clarity of images taken by different devices are different, which in turn brings problems of reduced recognition accuracy, pass rate and application security. Based on this, an embodiment of the present disclosure provides an image processing method, which can improve object recognition accuracy, pass rate and application security in practical applications.

The image processing method provided by the embodiments of the present disclosure can be applied to a terminal, where the terminal may include but not limited to access control equipment, attendance equipment, cash register equipment, etc., in some implementations, the above-mentioned access control equipment, attendance equipment, cash register equipment and other terminals A camera device is provided, wherein the camera device may be a camera device integrated with the terminal, or may be a separate camera device connected in a wired or wireless manner.

FIG. 1 is a flowchart of an image processing method provided by an embodiment of the present disclosure. As shown in FIG. 1, the image processing method is used in a terminal, and the image processing method includes steps S101 to S102, wherein:

Step S101: In response to a first object recognition instruction, acquire a first object image and a target reference image of a target object collected by a local device.

Wherein, the first object recognition instruction may be triggered based on different trigger operations in combination with different actual application requirements. In some implementations, the first object recognition instruction can be triggered based on the acquisition operation of the target object; for example, a button is set on the access control device to trigger the shooting of the user's face image, and correspondingly, the camera on the access control device can be triggered by pressing the button Execute the operation of collecting the user's face image. Correspondingly, when the camera device performs the operation of collecting the user's (face) image, the above-mentioned first object recognition instruction can be triggered, and then the local device can be called to collect the first object image of the target object. (including images of the target object), target reference images can also be acquired.

In some implementations, the target reference image is obtained by updating the registration image based on the local object image, wherein the registration image is an image of the target object collected by a remote device, and the registration image can be used for object registration. In some implementations, the local object image may be an object image captured by a local device. In some implementations, the target reference image may be obtained in advance, that is, obtained before the first object recognition instruction is triggered, and stored in a corresponding database or cache, after the first object recognition instruction is triggered, from the corresponding The database or cache reads the target reference image. In some implementations, the target reference image may serve as a reference image for identifying the target object. In some implementation manners, an object image satisfying a preset condition in the local object image may be selected to update the registration image to obtain the above-mentioned target object image. Correspondingly, the above-mentioned target reference image (that is, an object image satisfying a preset condition) may at least include one of the following:

The first object image selected from the local object image and the registration image is greater than or equal to the preset threshold;

The first object image selected from the local object image whose similarity with the registered image is greater than or equal to a preset threshold, and whose image quality analysis result satisfies the preset quality condition;

The similarity between the first selected local object image and the registered image is greater than or equal to a preset threshold, and when the local object image is collected, the orientation information of the target object relative to the camera device is at least one object image with specified orientation information;

The first object image selected from the local object image whose similarity with the registration image is greater than or equal to a preset threshold, and whose time attribute information when the local object image is collected matches the acquisition time attribute information of the registration image;

The first one selected from the local object image whose similarity with the registered image is greater than or equal to the preset threshold, and the time when the local object image is collected, the object image located in the target time period, and the target time period is the execution frequency of the preset operation The period of time greater than or equal to the preset frequency;

The object image with the highest similarity to the registered image is selected from the local object images;

Select the object image with the highest weighted average sum of the similarity with the registration image and the corresponding value of the living body detection result from the local object image.

It should be noted that the above-mentioned local device may be the above-mentioned terminal provided with the camera device integrated or separate from the terminal.

In some implementations, the target object can be different according to the actual application scenario. For example, in an application scenario that requires face recognition, the target object can be a user's face; for example, in an application scenario that requires iris recognition, The target object may be the iris of a certain user; for example, in an application scenario requiring fingerprint-based identification, the target object may be the fingerprint of a certain user. Correspondingly, different types of target objects may also use different types of camera devices to capture corresponding object images.

In some implementation manners, the above method may further include: a step of obtaining a target reference image in advance. Assuming that the above target reference image is the first object image selected from the local object images whose similarity with the registered image is greater than or equal to the preset threshold, correspondingly, Fig. 2 is a pre-obtained target reference image provided by the embodiment of the present disclosure As shown in FIG. 2 , obtaining the target reference image in advance may include steps S201 to S203:

Step S201: In response to a second object recognition instruction, acquire a second object image registration image of a target object collected by a local device.

Step S202: performing object recognition on the second object image based on the registered image to obtain a second object recognition result;

Step S203: When the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold, update the registration image based on the second object image to obtain a target reference image.

In some implementations, the local object image may be an object image captured by a local device. In some implementations, the above-mentioned second object recognition instruction may be an object recognition instruction triggered before the first object recognition instruction. When the second object recognition instruction is triggered, the reference image used as the recognition target object may be the above-mentioned registered image.

In some implementations, after the second object recognition instruction is triggered, in addition to performing the above acquisition of the second object image of the target object collected by the local device and the registration image of the target object, the registration image is updated based on the second object image to obtain the target reference Image operation; it is also possible to perform corresponding operations based on the second object recognition result in combination with actual application requirements, for example, when the second object recognition result indicates that the similarity between the second object image and the registered image is greater than or equal to a preset threshold , perform operations such as opening access control, entering attendance information, and making payment.

In the embodiment of the present disclosure, the second object recognition result may be the similarity between the registration image and the second object image. Correspondingly, the above step S202 may include calculating the similarity between the second object image and the registration image. In some implementations, the similarity between images may include but not limited to Euclidean distance, Manhattan distance, cosine distance, etc. between images.

In some implementations, the preset threshold can be set in combination with the requirements for object recognition accuracy and object recognition pass rate in practical applications. Among them, the higher the requirement for object recognition accuracy, the higher the preset threshold, and correspondingly, the relatively lower object recognition pass rate; Relatively higher.

In some implementations, when the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold, it may be determined that the object in the second object image is the target object, and correspondingly, the The registration image is updated based on the second object image to obtain a target reference image.

In the above embodiments, when the object image collected by the local device is recognized as the image of the target object, updating the preset object image collected by the non-local device based on the object image collected by the local device can effectively ensure the subsequent object recognition. Pass rate and application security.

In some implementations, when it is necessary to perform object recognition in combination with registered images collected by remote devices, optical property correction can be performed on registered images to reduce the difference in optical properties between images collected by local devices and images collected by remote devices Correspondingly, FIG. 3 is a schematic flowchart of performing object recognition on a second object image based on a registered image to obtain a second object recognition result provided by an embodiment of the present disclosure. As shown in FIG. 3 , the above step S202 may include step S301 To step S303:

Step S301: Perform optical attribute difference identification on the registration image and the second object image to obtain optical attribute difference information;

Step S302: Perform optical property correction on the registered image based on the optical property difference information to obtain a corrected image;

Step S303: performing object recognition on the second object image based on the corrected image to obtain a second object recognition result.

In some implementations, the optical property difference information may represent difference information in optical properties of the registered image relative to the second object image. Wherein, the optical properties may include but not limited to the definition, brightness, etc. of the image.

In some implementations, the optical property correction of the registered image can be performed in combination with histogram equalization, image distribution and the like.

In some implementations, step S303 may include: calculating a similarity between the second object image and the corrected image, and using the similarity as a specific refinement of the second object recognition result.

In the above embodiment, when it is necessary to perform object recognition in combination with registered images collected by remote devices, by correcting the optical properties of the registered images and the second object map collected by local devices, the optical properties of images caused by device differences can be effectively reduced. Differences, thereby improving the object recognition pass rate and application security.

In some embodiments, the updating of the registered image based on the second object image to obtain the target reference image includes step S2031, wherein:

Step S2031, using the second object image to replace the registered image to obtain a target reference image.

In some implementation manners, the registration image may be directly replaced by the second object image of the target object collected by the local device, and then the second object image may be used as a reference image for subsequent object recognition.

In the above embodiment, the object image collected by the local device is used to replace the preset object image collected by the non-local device, which can effectively avoid the difference in optical properties of the images collected by different devices, thereby improving the accuracy, pass rate and application of object recognition. security.

In some embodiments, the updating of the registered image based on the second object image to obtain the target reference image includes step S2032, wherein:

Step S2032, using the second object image and the registration image as target reference images.

In the above implementation manners, the object image collected by the local device is added to the reference image for object recognition, which can improve the object recognition accuracy, pass rate, and application security.

In some implementations, it is assumed that the above target reference image is the first object image selected from the local object images whose similarity with the registration image is greater than or equal to a preset threshold and whose image quality analysis result satisfies a preset quality condition. The step S203 may include steps S2041 to S2042, wherein:

Step S2041, when the second object recognition result indicates that the similarity between the second object image and the registered image is greater than or equal to a preset threshold, perform image quality analysis on the second object image to obtain an image quality analysis result;

Step S2042, if the image quality analysis result satisfies the preset quality condition, update the registered image based on the second object image to obtain the target reference image;

Among them, there may be one or more indicators that can be used to measure the image quality, and may be preset in combination with actual applications.

In some implementations, it is assumed that the definition of the image is used as an index to measure the image quality. Correspondingly, the preset quality condition may include a preset definition threshold, and the preset definition threshold may be combined with the target reference image in actual applications. Sharpness requires setting.

In some implementation manners, the aforementioned performing image quality analysis on the second object image to obtain an image quality analysis result may include: performing sharpness identification on the second object image to obtain image sharpness. Wherein, the definition recognition of the image may include but not limited to realize by combining the Laplacian algorithm. In some implementations, when the image sharpness of the second object image is greater than a preset sharpness threshold, it may be determined that the image quality analysis result satisfies a preset quality condition.

In some implementations, it is assumed that the number of objects in the image is used as an index to measure the image quality, and correspondingly, the preset quality condition may include a threshold of the number of objects, wherein, in a scene where a certain target object is recognized, the object image only includes The target object is more conducive to the accuracy of object recognition. In some implementations, the above object number threshold may be 1. Correspondingly, performing image quality analysis on the second object image to obtain an image quality analysis result may include performing object number recognition on the second object image to obtain the object number. In some implementations, object detection may be performed on the second object image based on a pre-trained object detection network to detect the number of objects in the second object image. In some implementations, when the number of objects corresponding to the second object image is 1 (a preset image quality index), it may be determined that the image quality analysis result satisfies the preset quality condition.

In the above embodiment, when the second object image collected by the local device is recognized as the image of the target object, combined with the image quality of the second object image, if the quality of the second object image satisfies the preset quality condition, Updating the preset object image collected by the non-local device based on the second object image can effectively improve the quality of the target reference image, thereby better improving object recognition accuracy, pass rate, and application security.

In some implementations, it is assumed that the above-mentioned target reference image is the first one selected from the local object image and the similarity with the registered image is greater than or equal to a preset threshold, and when the local object image is collected, the orientation information of the target object relative to the camera device Specify orientation information for at least one object image. The above step S203 may include steps S2051 to S2052, wherein:

Step S2051, when the second object recognition result indicates that the similarity between the second object image and the registered image is greater than or equal to a preset threshold, perform object orientation recognition on the second object image to obtain an object orientation recognition result;

Step S2052, if the object orientation recognition result satisfies at least one preset orientation condition, update the registered image based on the second object image to obtain a target reference image.

Wherein, the object orientation recognition result may represent the orientation information of the target object relative to the camera device when the second object image is captured. In some implementations, the at least one preset orientation condition may be at least one orientation information of the target object relative to the camera device when the preset target reference image is captured. For example, assuming that a certain preset orientation condition is that when the image is captured, the orientation information of the target object relative to the camera device is directly opposite (the target object is directly facing the camera device); correspondingly, the above-mentioned object orientation recognition result indicates that the second object image is collected , when the orientation information of the target object relative to the camera device is directly opposite, it may be determined that the object orientation recognition result satisfies at least one preset orientation condition.

In some implementation manners, the object orientation recognition may be performed on the second object image in combination with a pre-trained object orientation recognition network to obtain an object orientation recognition result.

In some implementation manners, multiple object images of different orientations collected by the local device may be selected as target reference images. For example, taking the target object as a human face as an example, it is assumed that a human face facing the left relative to the camera device, a face facing the right relative to the camera device, and a face image facing the camera device can be used as the target Reference image; correspondingly, when the object orientation recognition result indicates that the second object image is captured, the orientation information of the target object relative to the camera device is that the face is facing left relative to the camera device, the face is facing right relative to the camera device, or the face is facing right. In the case of a camera, the registered image may be updated based on the second object image to obtain a target reference image.

In the above embodiment, on the basis that the second object image collected by the local device is recognized as the image of the target object, the orientation information of the target object corresponding to at least one preset orientation condition relative to the camera device is combined to select the image collected in a specified orientation. The registration image can be updated by using the object image, which can effectively ensure the quality of the collected object image, and further improve the object recognition accuracy, pass rate and application security.

In some implementations, it is assumed that the above-mentioned target reference image is the first one selected from the local object image, and the similarity with the registration image is greater than or equal to the preset threshold, and the time attribute information of the local object image is collected, and the collection of the registration image The temporal attribute information matches the object image. The above step S203 may include steps S2061 to S2062, wherein:

Step S2061, when the second object recognition result indicates that the similarity between the second object image and the registered image is greater than or equal to a preset threshold, acquire acquisition time attribute information of the registered image;

Step S2062, if the current time attribute information matches the acquisition time attribute information, update the registered image based on the second object image to obtain a target reference image.

Wherein, the acquisition time attribute information may be information capable of characterizing the time at which the registration image is acquired. The collection time attribute information may include but not limited to collection season, collection time period (such as 9:00-11:00 am in a day) and so on.

In some implementations, in order to better ensure the similarity between the reference image used to identify the target object and the registration image in the object image collected by the local device, the acquisition time attribute information of the registration image may be combined. For example, the acquisition time attribute information of the initial reference object image is summer, and if the current time attribute information indicates that the current season is also summer, it can be determined that the current time attribute information matches the acquisition time attribute information of the registered image, and then the update based on the second object image is performed. Register the image and get the operation of the target reference image.

In the above embodiments, the object images collected by the local device are screened in combination with the matching between the collection time attribute information of the registered image and the current time attribute information, which can better ensure that the object images collected by the local device are used as identification targets. The similarity between the reference image of the object and the registration image can be used to improve the accuracy of object recognition, pass rate and application security.

In some implementations, it is assumed that the above-mentioned target reference image is the first one selected from the local object image whose similarity with the registration image is greater than or equal to the preset threshold, and the time when the local object image is collected, the object image located in the target time period . The above step S203 may include steps S2071 to S2073, wherein:

Step S2071, when the second object recognition result indicates that the similarity between the second object image and the registered image is greater than or equal to the preset threshold, obtain the historical operation trigger time corresponding to the target object, and the historical operation trigger time is based on the target object corresponding The trigger time of the preset operation triggered by the object recognition result;

Step S2072, determine the target time period based on the historical operation trigger time, the target time period is the time period in which the execution frequency of the preset operation is greater than or equal to the preset frequency;

Step S2073, if the current time is within the target time period, update the registered image based on the second object image to obtain the target reference image.

Wherein, the historical operation trigger time is the trigger time when the preset operation is triggered based on the object recognition result corresponding to the target object; the above-mentioned target time period is the time period when the execution frequency of the preset operation is greater than or equal to the preset frequency. In some implementations, the preset frequency can be preset in combination with actual application scenarios. The time period can be a certain period of time in the preset cycle, for example, the preset cycle can be from 0:00 to 24:00 every day; it can also be from 0:00 on Monday to 24:00 on Sunday every week. Specifically, it can be combined actual application settings.

In practical applications, for example, in the application scenario of attendance check-in based on face recognition, the user triggers the operation of entering attendance information in a fixed time period of the day. Assuming that combined with the historical operation trigger time, it is determined every morning from 7:30 to 9 :00, and every afternoon from 17:00 to 18:30 is the target time period, then from 7:30 to 9:00 every morning, and every afternoon from 17:00 to 18:30, you can base on the second object The image is updated to register the image to obtain the target reference image.

In the above-mentioned embodiment, considering whether the collection time period of the object image is within the target time period, it is convenient to select the target reference of the target object collected by the local device during the peak time period when the preset operation is triggered according to the user’s habit of triggering the preset operation. image, thereby effectively reducing the difference between the subsequent target reference image and the object image to be recognized due to the different acquisition time, and improving the object recognition accuracy, pass rate and application security.

In some implementations, it is assumed that the above target reference image may be the object image selected from the local object images with the highest similarity with the registered image. FIG. As shown in Figure 4, the above method may also include steps S401 to S404, wherein:

Step S401: In response to the third object recognition instruction, acquire the third object image and the registration image of the target object collected by the local device;

Step S402: performing object recognition on the third object image based on the registered image to obtain a third object recognition result;

Step S403: Comparing the third object recognition result with the second object recognition result to obtain a comparison result;

Step S404: When the comparison result indicates that the similarity between the third object image and the registered image is greater than the similarity between the second object image and the registered image, update the target reference image based on the third object image.

Wherein, the above-mentioned third object recognition instruction may be an object recognition instruction triggered after the second object recognition instruction. The object recognition result may be a degree of similarity between the registration image and the third object image.

In some implementations, in order to better improve the quality of the reference image for identifying the target object, an object image with a higher degree of similarity to the registered image may be selected from the object images collected by the local device to continuously update the target reference image.

In practical applications, after the third object recognition instruction is triggered, in addition to performing the above-mentioned acquisition of the third object image and registration image of the target object captured by the local device, the operation of updating the target reference image based on the third object image; it can also be combined with the actual Application requirements, perform object recognition on the third object image based on the target reference image, and perform corresponding operations based on the object recognition result, for example, when the object recognition result indicates that the similarity between the third object image and the target reference image is greater than or equal to the preset When the threshold is exceeded, perform operations such as opening access control and entering attendance information.

In some implementation manners, the above-mentioned object recognition is performed on the third object image based on the registered image to obtain a specific refinement of the third object recognition result, and reference may be made to the specific refinement of the above step S202.

In the above embodiments, the target reference image is continuously updated by selecting an object image with a higher degree of similarity with the registered image from the object images collected by the local device, which can better improve the quality of the reference image for identifying the target object.

In some implementations, it is assumed that the above-mentioned target reference image is the object image selected from the local object image with the highest sum of the weighted average of the similarity with the registration image and the corresponding value of the living body detection result. A schematic flow chart of obtaining a target reference image in advance, as shown in FIG. 5 , the above method may further include steps S501 to S504, wherein:

Step S501: Obtain the first target image of the target object collected by the local device within a preset time period, and the fourth object recognition result between the first target image and the registered image;

Step S502: Perform liveness detection on the first target image to obtain a liveness detection result;

Step S503: Based on the fourth object recognition result and the living body detection result, filter out the second target image from the first target image;

Step S504: Update the registered image based on the second target image to obtain a target reference image.

Wherein, the first target image may be an object image whose matching degree with the initial reference image is greater than or equal to a preset threshold among images of the target object collected by the local device within a preset time period. In some implementations, the preset time period may be a preset update period of the registered image. The fourth object recognition result may be the similarity between the registration image and the first target image.

In some implementations, the above step S502 may include: performing life detection on the first target image based on a pre-trained life detection network to obtain a life detection result.

Wherein, the living body detection result may represent information about whether there is a living body (for example, a real person) when the first target image is captured. In practical applications, the living body detection result may be the probability that there is a living body when the first target image is collected.

In some implementations, the above step S503 may include: performing a weighted average or correlation of the value (similarity) corresponding to the object recognition result of any image in the first target image and the value (probability) corresponding to the living body detection result of the image. Adding, selecting the image with the highest weighted average or added value as the second target image, and then updating the registered image based on the second target image to obtain a target reference image. Wherein, for the specific detailed steps of the above step S504, please refer to the detailed detailed steps of the above step S203.

In addition, it should be noted that the respective weights of the fourth object recognition result and the living body detection result can be set according to actual application requirements.

In some implementations, the first of the locally collected object images, the sum of the values corresponding to the object recognition result and the living body detection result, or the object image whose weighted average value is greater than or equal to the preset value can also be updated. Register the image to get the target reference image.

In the above embodiments, the target reference image is selected in combination with the object recognition results and living body detection results of the object images collected by the local device within a period of time, which can better ensure that images of real objects are collected, and thus better ensure image quality.

In some implementations, the above method may also include steps S601 to S602:

Step S601, based on the preset update frequency, acquire the third target image of the target object collected by the local device in the current update period;

Step S602, updating the target reference image based on the image satisfying the preset condition in the third target image.

Wherein, the preset update frequency may be preset according to actual application requirements, and the preset update frequency may be the update frequency of the target reference image. In some implementations, the third target image may be an image of the target object collected by the local device in the current update period. For specific refinement of the images satisfying the preset conditions in the third target image, refer to the specific refinement of the target images satisfying the preset conditions when determining the target reference image above.

In the above embodiments, the target reference image is continuously updated according to the preset update frequency, which can better ensure the similarity between the target reference image and the target object, and effectively improve the image quality of the target reference image.

In some embodiments, the above step S602 may include step S6021 or step S6022, wherein:

Step S6021, replacing the target reference image with the target image satisfying the preset condition in the third target image;

Step S6022, adding object images satisfying preset conditions in the third target image to the target reference image.

In some implementation manners, the third target image of the target object collected by the local device in the current update period can directly replace the original target reference image, so as to update the target reference image.

In some implementations, the third target image of the target object collected by the local device in the current update period meets the preset conditions and can be added to the target reference image to update the target reference image; in some implementations , the upper limit of the number of images of the target reference image can also be set in advance, and when the number of images of the target reference image reaches the above upper limit of the number of images, the operation of adding object images satisfying the preset conditions in the third target image to the target reference image can be stopped.

In the above embodiment, in the process of updating the target reference image by using the object image collected by the local device in the current update period, the target reference image can be replaced or the object image that meets the preset condition in the current update period can be added to the target reference image. The update method of the target reference image can effectively improve the quality of the target reference image and at the same time increase the diversity of update methods of the target reference image.

In addition, it should be noted that the preset conditions listed above are examples. In practical applications, more preset conditions can be set in combination with actual application requirements, or at least two examples of different preset conditions can be randomly selected. Examples are combined into new presets and more.

In some implementations, when the target reference image includes multiple images, the above method may further include steps S701 to S702, wherein:

Step S701, displaying multiple images;

Step S702, in response to triggering the confirmation instruction based on at least one image in the plurality of images, updating the target reference image based on at least one image corresponding to the confirmation instruction.

In practical applications, in order to better ensure the quality of the target reference image, the target reference image may be displayed to the user so that the user can select and confirm the target reference image.

In some implementations, when the target reference image includes one image, the target reference image may also be presented for confirmation by the user. Correspondingly, if the confirmation instruction is triggered based on the displayed target reference image, the corresponding The image is used as the target reference image; otherwise, the target reference image can be obtained by recombining the preset conditions.

Step S102: Perform object recognition on the first object image based on the target reference image to obtain a first object recognition result.

In some implementations, the first object recognition result may be the similarity between the target reference image and the first object image.

A target reference image may include one or more images. In the case where the target reference image includes multiple images, the above step S102 may include steps S1021 to S1022, wherein:

Step S1021, performing object recognition on the first object image based on the plurality of images, respectively, to obtain a plurality of sub-object recognition results;

Step S1022, generating a first object recognition result based on the multiple sub-object recognition results.

In some embodiments, multiple sub-object recognition results can be added to obtain the above-mentioned first object recognition result; weight information corresponding to multiple images can also be preset, and weighted average of multiple sub-object recognition results can be obtained based on the weight information. The above first object recognition result.

In some implementations, when the target reference image includes the above-mentioned second object image and the registration image, the above-mentioned step S102 may include steps S1031 to S1033, wherein:

Step S1031, performing object recognition on the first object image based on the registered image to obtain a first sub-object recognition result;

Step S1032, performing object recognition on the first object image based on the second object image to obtain a second sub-object recognition result;

Step S1033: Determine the first object recognition result according to the first sub-object recognition result and the second sub-object recognition result.

In some implementation manners, for details of step S1031 and step S1032, refer to the details of step S202 above.

In some implementations, step S1033 may include adding the first sub-object recognition result and the second sub-object recognition result to obtain the above-mentioned first sub-object recognition result; it may also be based on preset weight information, based on the weight information for the first sub-object The object recognition result and the second sub-object recognition result are weighted and averaged to obtain the above-mentioned first object recognition result.

In the above embodiments, in the case that the target reference image includes multiple images, object recognition is performed on the first object image based on the multiple target reference images, which can better ensure the accuracy of object recognition.

In some implementations, when the first object recognition result indicates that the similarity between the first object image and the target reference image is greater than or equal to a preset threshold, a preset operation may be performed. Wherein, the preset operation may be different according to different actual application requirements. For example, the preset operation may include but not limited to opening the access control, entering attendance information, payment and other operations. In some implementations, when the first object recognition result indicates that the similarity between the first object image and the target reference image is less than a preset threshold, preset prompt information may be fed back so as to prompt the user to retake the image of the target object.

It can be seen from the above technical solutions provided by the embodiments of the present disclosure that in the present disclosure, when the object image collected by the local device is recognized as the image of the target object, updating the registered image collected by the remote device based on the object image collected by the local device can Effectively guarantee the subsequent object recognition pass rate and application security.

Those skilled in the art can understand that in the above method of specific implementation, the writing order of each step does not mean a strict execution order and constitutes any limitation on the implementation process. The specific execution order of each step should be based on its function and possible The inner logic is OK.

FIG. 6 is a schematic diagram of the composition and structure of an image processing device provided by an embodiment of the present disclosure. Referring to Figure 6, the device includes:

The first image acquiring part 610 is configured to execute in response to the first object recognition instruction, acquire the first object image and the target reference image of the target object collected by the local device, and the target reference image is obtained by updating the registered image based on the local object image , the registered image is the image of the target object collected by the remote device, and the registered image is used for object registration;

The first object recognition part 620 is configured to perform object recognition on the first object image based on the target reference image to obtain a first object recognition result.

In some embodiments, the target reference image includes at least one of the following:

The first object image selected from the local object image and the registration image is greater than or equal to the preset threshold;

The first object image selected from the local object image whose similarity with the registered image is greater than or equal to a preset threshold, and whose image quality analysis result satisfies the preset quality condition;

The similarity between the first selected local object image and the registered image is greater than or equal to a preset threshold, and when the local object image is collected, the orientation information of the target object relative to the camera device is at least one object image with specified orientation information;

The first object image selected from the local object image whose similarity with the registration image is greater than or equal to a preset threshold, and whose time attribute information when the local object image is collected matches the acquisition time attribute information of the registration image;

The first one selected from the local object image whose similarity with the registered image is greater than or equal to the preset threshold, and the time when the local object image is collected, the object image located in the target time period, and the target time period is the execution frequency of the preset operation The period of time greater than or equal to the preset frequency;

The object image with the highest similarity to the registered image is selected from the local object images;

Select the object image with the highest weighted average sum of the similarity with the registration image and the corresponding value of the living body detection result from the local object image.

In some embodiments, the above-mentioned device also includes:

The second image acquiring part is configured to acquire a second object image and a registered image of the target object captured by the local device in response to a second object recognition instruction, where the second object recognition instruction is an object triggered before the first object recognition instruction identification instructions;

The second object recognition part is configured to perform object recognition on the second object image based on the registered image to obtain a second object recognition result;

The first registration image updating part is configured to execute updating the registration image based on the second object image to obtain the target reference when the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold. image.

In some embodiments, the second object recognition component includes:

The optical attribute difference identification part is configured to perform optical attribute difference identification on the registration image and the second object image, and obtain optical attribute difference information;

The optical attribute correction part is configured to perform optical attribute correction on the registered image based on the optical attribute difference information to obtain a corrected image;

The object recognition part is configured to perform object recognition on the second object image based on the corrected image to obtain a second object recognition result.

In some embodiments, the first registration image update section includes:

The first registered image updating subpart is configured to replace the registered image with the second object image to obtain the target reference image;

or,

The second registered image updating subsection is configured to perform using the second object image and the registered image as target reference images.

In some embodiments, the first registration image update section includes:

The image quality analysis part is configured to perform image quality analysis on the second object image when the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold value, to obtain an image quality analysis result;

The third registered image updating subpart is configured to update the registered image based on the second object image to obtain the target reference image when the image quality analysis result meets the preset quality condition;

or,

The object orientation recognition part is configured to perform object orientation recognition on the second object image when the second object recognition result indicates that the similarity between the second object image and the registered image is greater than or equal to a preset threshold, to obtain the object orientation recognition result;

The fourth registered image updating subsection is configured to update the registered image based on the second object image to obtain a target reference image when the object orientation recognition result meets at least one preset orientation condition;

or,

The acquisition time attribute acquisition part is configured to acquire the acquisition time attribute information of the registration image when the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold;

The fifth registration image updating subpart is configured to update the registration image based on the second object image to obtain a target reference image when the current time attribute information matches the acquisition time attribute information.

or,

The historical operation trigger time acquisition part is configured to acquire the historical operation trigger time corresponding to the target object when the second object recognition result indicates that the similarity between the second object image and the registered image is greater than or equal to a preset threshold, and the historical operation trigger time The trigger time is the trigger time when the preset operation is triggered based on the object recognition result corresponding to the target object;

The target time period determining part is configured to determine the target time period based on the historical operation trigger time, and the target time period is a time period in which the execution frequency of the preset operation is greater than or equal to the preset frequency;

The sixth registered image updating subpart is configured to update the registered image based on the second object image to obtain the target reference image when the current time is within the target time period.

In some implementations, the above apparatus further includes: a third image acquisition part configured to acquire a third object image and a registered image of the target object captured by the local device in response to the third object recognition instruction; the third object recognition part, configured to perform object recognition on the third object image based on the registered image to obtain a third object recognition result; the object recognition result comparison part is configured to perform a comparison between the third object recognition result and the second object recognition result to obtain a comparison result; The first target reference image updating section configured to perform updating the target based on the third target image in a case where the comparison result indicates that the similarity between the third target image and the registered image is greater than the similarity between the second target image and the registered image. Reference image.

In some embodiments, the above-mentioned device also includes:

The data acquisition part is configured to execute the acquisition of the first target image of the target object collected by the local device within a preset time period, and the fourth object recognition result between the first target image and the registration image of the target object; the living body detection part, It is configured to perform liveness detection on the first target image to obtain a liveness detection result; the target image screening part is configured to perform screening of the second target image from the first target image based on the fourth object recognition result and the liveness detection result ; The second registered image updating part is configured to update the registered image based on the second target image to obtain a target reference image.

With regard to the apparatus in the above embodiments, the specific manner in which each part executes operations has been described in detail in the embodiments related to the method, and will not be described in detail here.

In the embodiments of the present disclosure and other embodiments, a "part" may be a part of a circuit, a part of a processor, a part of a program or software, etc., of course it may also be a unit, a module or a non-modular one.

FIG. 7 is a schematic diagram of a hardware structure of an electronic device 700 provided by an embodiment of the present disclosure. The electronic device 700 may be a terminal, and its internal structure may be as shown in FIG. 7 . The electronic device 700 includes a processor 711 connected through a system bus 710 , a memory, a network interface 713 , a display screen 714 and an input device 715 . Wherein, the processor 711 of the electronic device is used to provide calculation and control capabilities. The memory of the electronic device may include, but not limited to, a non-volatile storage medium 7121, an internal memory 7122, and the like. The non-volatile storage medium stores an operating system 716 and a computer program 717 . The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface 713 of the electronic device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor 711, the above image processing method can be realized. The display screen 714 of the electronic device may include but not limited to a liquid crystal display screen or an electronic ink display screen, etc., and the input device 715 of the electronic device may be a touch layer covered on the display screen, or a key set on the casing of the electronic device, A trackball or a touchpad, or an external keyboard, touchpad, or mouse.

Those skilled in the art can understand that the structure shown in FIG. 7 is a block diagram of a partial structure related to the disclosed solution, and does not constitute a limitation on the electronic equipment to which the disclosed solution is applied. The specific electronic equipment may include There may be more or fewer components than shown in the figures, or certain components may be combined, or have different component arrangements.

An embodiment of the present disclosure also provides an electronic device, including: a processor; a memory for storing an instruction executable by the processor; wherein the processor is configured to execute the instruction to realize the image in the embodiment of the present disclosure Approach.

The embodiment of the present disclosure also provides a storage medium, and when the instructions in the storage medium are executed by the processor of the electronic device, the electronic device can execute the image processing method in the embodiment of the present disclosure.

The embodiment of the present disclosure also provides a computer program product, the computer program product includes a computer program or an instruction, and when the computer program or instruction is run on the electronic device, the electronic device executes the image processing method in the embodiment of the present disclosure.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be realized by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium , when the computer program is executed, it may include the procedures of the embodiments of the above-mentioned methods. Wherein, any reference to memory, storage, database or other media used in various embodiments provided by the present disclosure may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (Read Only Memory, ROM), programmable ROM (Programmable Read Only Memory, PROM), electrically programmable ROM (Erasable Programmable Read Only Memory, EPROM), electrically erasable programmable ROM (Electrically Erasable Programmable Read Only Memory, EEPROM) or flash memory. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as Static Random Access Memory (SRAM), Dynamic RAM (Dynamic Random Access Memory, DRAM), Synchronous DRAM (Synchronous Dynamic Random Access Memory, SDRAM), dual Data rate SDRAM (Double Data Rate Synchronous Dynamic Random Access Memory, DDRSDRAM), enhanced SDRAM (Enhanced Synchronous Dynamic Random Access Memory, ESDRAM), synchronous link DRAM (Synchlink Dynamic Random Access Memory, SLDRAM), memory bus direct RAM (Rambus Dynamic Random Access Memory, RDRAM), and direct memory bus dynamic RAM (Direct Rambus Dynamic Random Access Memory, DRDRAM), etc.

Other embodiments of the present disclosure will be readily apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. The present disclosure is intended to cover any modification, use or adaptation of the present disclosure. These modifications, uses or adaptations follow the general principles of the present disclosure and include common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. . The specification and examples are to be considered exemplary only, with the true scope and spirit of the disclosure indicated by the appended claims.

It should be understood that the present disclosure is not limited to the precise constructions which have been described above and shown in the drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Industrial Applicability

Embodiments of the present disclosure provide an image processing method and device, electronic equipment, a storage medium, and a computer program product. The method includes: in response to a first object recognition instruction, acquiring a first object image and a target reference of a target object collected by a local device image, the target reference image is obtained by updating the registered image based on the local object image, the registered image is the image of the target object collected by the remote device, and the registered image is used for object registration; the object recognition is performed on the first object image based on the target reference image, A first object recognition result is obtained. The above solution can improve the accuracy rate of object recognition, the pass rate and the security of the application.

Claims (19)

1. An image processing method, comprising:

   In response to the first object recognition instruction, acquire the first object image and the target reference image of the target object collected by the local device, the target reference image is obtained by updating the registered image based on the local object image, and the registered image is collected by the remote device an image of the target object, the registration image is used for object registration;

   Perform object recognition on the first object image based on the target reference image to obtain a first object recognition result.
2. The image processing method according to claim 1, wherein the target reference image comprises at least one of the following:

   The first object image selected from the local object images whose similarity with the registration image is greater than or equal to a preset threshold;

   Selecting the first object image from the local object image whose similarity with the registered image is greater than or equal to a preset threshold and whose image quality analysis result meets a preset quality condition;

   The similarity between the first one selected from the local object image and the registration image is greater than or equal to a preset threshold, and when the local object image is collected, the orientation information of the target object relative to the camera device is at least one An object image specifying orientation information;

   The similarity between the first one selected from the local object image and the registration image is greater than or equal to a preset threshold, and the time attribute information when the local object image is collected is the same as the collection time attribute information of the registration image matching object image;

   The similarity between the first one selected from the local object image and the registration image is greater than or equal to a preset threshold, and the time when the local object image is collected, the object image located in the target time period, the target time The period is the time period during which the execution frequency of the preset operation is greater than or equal to the preset frequency;

   An object image with the highest similarity to the registered image selected from the local object images;

   An object image with the highest weighted average sum of the similarity with the registration image and the corresponding value of the living body detection result is selected from the local object image.
3. The image processing method according to claim 1 or 2, wherein the method further comprises:

   Acquire a second object image of the target object collected by the local device and the registration image in response to a second object recognition instruction, where the second object recognition instruction is an object triggered before the first object recognition instruction identify instructions;

   performing object recognition on the second object image based on the registered image to obtain a second object recognition result;

   When the second object recognition result indicates that the similarity between the second object image and the registered image is greater than or equal to a preset threshold, update the registered image based on the second object image to obtain the target Reference image.
4. The image processing method according to claim 3, wherein performing object recognition on the second object image based on the registered image to obtain a second object recognition result comprises:

   Performing optical attribute difference identification on the registration image and the second object image to obtain optical attribute difference information;

   performing optical property correction on the registered image based on the optical property difference information to obtain a corrected image;

   Perform object recognition on the second object image based on the corrected image to obtain the second object recognition result.
5. The image processing method according to claim 3, wherein said updating said registration image based on said second object image to obtain said target reference image comprises:

   replacing the registration image with the second object image to obtain the target reference image;

   or,

   The second object image and the registration image are used as the target reference image.
6. The image processing method according to claim 3, wherein, when the second object recognition result indicates that the similarity between the second object image and the registered image is greater than or equal to a preset threshold, based on the Updating the registration image with the second object image to obtain the target reference image, including:

   When the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold, perform image quality analysis on the second object image to obtain an image quality analysis result ;

   When the image quality analysis result meets a preset quality condition, updating the registration image based on the second object image to obtain the target reference image;

   or,

   When the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold, perform object orientation recognition on the second object image to obtain an object orientation recognition result ;

   In a case where the object orientation recognition result satisfies at least one preset orientation condition, updating the registration image based on the second object image to obtain the target reference image;

   or,

   If the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold, acquiring acquisition time attribute information of the registration image;

   In the case that the current time attribute information matches the collection time attribute information, updating the registration image based on the second object image to obtain the target reference image;

   or,

   When the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold, acquire the historical operation trigger time corresponding to the target object, and the historical operation trigger The time is a trigger time for triggering the preset operation based on the object recognition result corresponding to the target object;

   determining a target time period based on the historical operation trigger time, where the target time period is a time period in which the execution frequency of the preset operation is greater than or equal to a preset frequency;

   When the current time is within the target time period, the registration image is updated based on the second object image to obtain the target reference image.
7. The image processing method according to claim 3, wherein said method further comprises:

   In response to a third object recognition instruction, acquire a third object image and the registration image of the target object captured by the local device;

   performing object recognition on the third object image based on the registered image to obtain a third object recognition result;

   comparing the third object recognition result with the second object recognition result to obtain a comparison result;

   In a case where the comparison result indicates that the similarity between the third object image and the registration image is greater than the similarity between the second object image and the registration image, updating the third object image based on the third object image the target reference image.
8. The image processing method according to claim 1 or 2, wherein the method further comprises:

   Acquiring a first target image of the target object captured by the local device within a preset time period, and a fourth object recognition result between the first target image and the registration image;

   performing a living body detection on the first target image to obtain a living body detection result;

   filtering out a second target image from the first target image based on the fourth object recognition result and the living body detection result;

   The registration image is updated based on the second target image to obtain the target reference image.
9. An image processing device, comprising:

   The first image acquisition part is configured to execute in response to the first object recognition instruction, acquire the first object image and the target reference image of the target object captured by the local device, and the target reference image is obtained by updating the registration image based on the local object image Wherein, the registration image is an image of the target object collected by a remote device, and the registration image is used for object registration;

   The first object recognition part is configured to perform object recognition on the first object image based on the target reference image to obtain a first object recognition result.
10. The image processing device according to claim 9, wherein the target reference image comprises at least one of the following:

    The first object image selected from the local object images whose similarity with the registration image is greater than or equal to a preset threshold;

    Selecting the first object image from the local object image whose similarity with the registered image is greater than or equal to a preset threshold and whose image quality analysis result meets a preset quality condition;

    The similarity between the first one selected from the local object image and the registration image is greater than or equal to a preset threshold, and when the local object image is collected, the orientation information of the target object relative to the camera device is at least one An object image specifying orientation information;

    The similarity between the first one selected from the local object image and the registration image is greater than or equal to a preset threshold, and the time attribute information when the local object image is collected is the same as the collection time attribute information of the registration image matching object image;

    The similarity between the first one selected from the local object image and the registration image is greater than or equal to a preset threshold, and the time when the local object image is collected, the object image located in the target time period, the target time The period is the time period during which the execution frequency of the preset operation is greater than or equal to the preset frequency;

    An object image with the highest similarity to the registered image selected from the local object images;

    An object image with the highest weighted average sum of the similarity with the registration image and the corresponding value of the living body detection result is selected from the local object image.
11. The image processing device according to claim 9 or 10, wherein the device further comprises:

    The second image acquisition part is configured to acquire the second object image of the target object captured by the local device and the registration image in response to a second object recognition instruction, the second object recognition instruction is in the an object recognition command triggered before the first object recognition command;

    The second object recognition part is configured to perform object recognition on the second object image based on the registration image to obtain a second object recognition result;

    The first registration image updating part is configured to perform, in the case that the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold, based on the second object and image updating the registered image to obtain the target reference image.
12. The image processing apparatus according to claim 11, wherein the second object recognition part further comprises:

    The optical attribute difference identification part is configured to perform optical attribute difference identification on the registration image and the second object image, and obtain optical attribute difference information;

    The optical property correction part is configured to perform optical property correction on the registration image based on the optical property difference information, to obtain a corrected image;

    The object recognition part is configured to perform object recognition on the second object image based on the corrected image to obtain the second object recognition result.
13. The image processing apparatus according to claim 11, wherein said first registered image updating part further comprises:

    The first registration image updating subpart is configured to replace the registration image with the second object image to obtain the target reference image;

    or,

    The second registered image updating subsection is configured to perform using the second object image and the registered image as the target reference image.
14. The image processing apparatus according to claim 11, wherein said first registered image updating part further comprises:

    The image quality analysis part is configured to perform, in the case that the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold, perform Image quality analysis to obtain image quality analysis results;

    The third registration image updating subsection is configured to update the registration image based on the second object image to obtain the target reference image when the image quality analysis result satisfies a preset quality condition;

    or,

    The object orientation recognition part is configured to perform, in the case that the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold, perform Object orientation recognition, get the object orientation recognition result;

    The fourth registration image updating subsection is configured to perform updating the registration image based on the second object image to obtain the target reference image when the object orientation recognition result satisfies at least one preset orientation condition;

    or,

    An acquisition time attribute acquisition section configured to perform acquisition of the registration image when the second object recognition result indicates that the similarity between the second object image and the registration image is greater than or equal to a preset threshold Time attribute information;

    The fifth registration image updating subpart is configured to update the registration image based on the second object image to obtain the target reference image when the current time attribute information matches the collection time attribute information;

    or,

    The historical operation trigger time acquiring part is configured to acquire the target object corresponding The historical operation trigger time, the historical operation trigger time is the trigger time for triggering the preset operation based on the object recognition result corresponding to the target object;

    The target time period determining part is configured to determine a target time period based on the historical operation trigger time, the target time period being a time period in which the execution frequency of the preset operation is greater than or equal to a preset frequency;

    The sixth registration image updating subpart is configured to update the registration image based on the second object image to obtain the target reference image when the current time is within the target time period.
15. The image processing device according to claim 11, wherein the device further comprises:

    The third image acquisition part is configured to execute in response to a third object recognition instruction, acquire a third object image and the registration image of the target object captured by the local device;

    The third object recognition part is configured to perform object recognition on the third object image based on the registered image to obtain a third object recognition result;

    The object recognition result comparison part is configured to compare the third object recognition result with the second object recognition result to obtain a comparison result;

    A first target reference image updating section configured to perform a process in which the comparison result indicates that the degree of similarity between the third object image and the registration image is greater than the degree of similarity between the second object image and the registration image. In some cases, the target reference image is updated based on the third object image.
16. The image processing device according to claim 9 or 10, wherein the device further comprises:

    A data acquisition part configured to acquire the first target image of the target object collected by the local device within a preset time period, and a fourth object recognition result between the first target image and the registration image ;

    The live body detection part is configured to perform live body detection on the first target image to obtain a live body detection result;

    a target image screening part configured to perform screening of a second target image from the first target image based on the fourth object recognition result and the living body detection result;

    The second registered image updating part is configured to update the registered image based on the second target image to obtain the target reference image.
17. An electronic device comprising:

    processor;

    memory for storing said processor-executable instructions;

    Wherein, the processor is configured to execute the instructions, so as to realize the image processing method according to any one of claims 1-8.
18. A computer-readable storage medium, when the instructions in the storage medium are executed by the processor of the electronic device, the image processing device can execute the image processing method according to any one of claims 1 to 8.
19. A computer program product, the computer program product comprising a computer program or an instruction, when the computer program or instruction is run on an electronic device, the electronic device is made to execute any one of claims 1 to 8. Methods.

Images