CN118155240A - Identity recognition method and device and electronic equipment - Google Patents

Abstract

The application discloses an identity identification method, an identity identification device and electronic equipment. The method comprises the following steps: a sequence of images presenting a target object is acquired. And evaluating the images in the image sequence by taking the characteristic presentation quality of the target object as a standard. And selecting a first preset number of representative images from the image sequence according to the evaluation result. Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database. The application can solve the problem of low identification efficiency in the existing cross-mirror tracking technology.

Description

Identity recognition method and device and electronic equipment

Technical Field

The present application relates to the field of intelligent security technologies, and in particular, to an identity identification method, an identity identification device, and an electronic device.

Background

With the rapid development of internet technology, the field of intelligent security and protection is rapidly rising, and the information processing and analysis of human images in videos and pedestrian interaction behaviors become important points of attention.

Cross-mirror tracking is an important application of intelligent security. The research object of cross-mirror tracking is the integral characteristics of pedestrians, including clothing, body type, hairstyle, gait, gesture and the like, and aims to realize the identification of pedestrians. While existing face recognition techniques help achieve this goal, because pedestrians are not always facing the camera lens under different cameras and scenes, a cross-mirror tracking technique is required as an important complement to face recognition techniques.

The existing cross-mirror tracking system generally performs warehouse-in retrieval comparison on direct pedestrian tracking images, then performs feature similarity calculation on query images and a retrieved base, and finally screens identity information according to the sequence of feature similarity. The quality of the pedestrian tracking image is good or bad, and the retrieved base is too large when the quality is poor, so that the subsequent identification efficiency is affected.

Disclosure of Invention

The embodiment of the application provides an identity identification method, an identity identification device and electronic equipment, which solve the problem of low identity identification efficiency in the existing cross-mirror tracking technology.

In order to solve the technical problems, the embodiment of the application is realized as follows:

In a first aspect, an identification method is provided, including:

acquiring an image sequence presenting a target object;

Evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard;

Selecting a first preset number of representative images from the image sequence according to the evaluation result;

Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

In a second aspect, an identification device is provided, comprising:

the acquisition module acquires an image sequence presenting a target object;

The evaluation module is used for evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard;

The selecting module is used for selecting a first preset number of representative images from the image sequence according to the evaluation result;

The comparison module is used for importing the characteristic information of the target object in the representative image into a database for comparison so as to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

In a third aspect, an electronic device is provided, comprising: a processor; and a memory arranged to store computer executable instructions that, when executed, cause the processor to:

acquiring an image sequence presenting a target object;

Evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard;

Selecting a first preset number of representative images from the image sequence according to the evaluation result;

Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

In a fourth aspect, a computer program product is presented, the computer program product storing one or more programs that, when executed by an electronic device comprising a plurality of application programs, cause the electronic device to:

acquiring an image sequence presenting a target object;

Evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard;

Selecting a first preset number of representative images from the image sequence according to the evaluation result;

Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

The application can be applied to identity recognition based on a cross-mirror tracking technology. Specifically, when an image sequence presenting a target object is obtained, evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard, so that a representative image with higher characteristic presentation quality is selected in the image sequence according to an evaluation result; and then, the characteristic information of the target object in the representative image is imported into a database to be searched and compared with the characteristic information of the base, and the base obtained by searching and comparing is more accurate due to higher characteristic presentation quality of the representative image, and finally, the relevant calculation of the identity recognition is carried out in the base with a small range, so that the efficiency of the identity recognition can be greatly improved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:

fig. 1 is a schematic flow chart of a first identification method according to an embodiment of the present application.

Fig. 2 is a schematic diagram of a second flow chart of an identification method according to an embodiment of the application.

Fig. 3 is a schematic structural diagram of an identification device according to an embodiment of the present application.

Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

As described above, the conventional cross-mirror tracking system generally performs the warehouse entry detection on the pedestrian tracking image, so that the quality of the pedestrian tracking image will directly determine the size of the search base, thereby affecting the efficiency of identity recognition.

Therefore, the application aims to provide an identity recognition scheme based on a cross-mirror tracking technology, which can evaluate the tracked pedestrian tracking image by taking the characteristic presentation quality as a standard, so that the high-quality pedestrian tracking image is screened for warehousing retrieval comparison, and the identity recognition rate is accelerated while the retrieval precision is improved.

For the purposes, technical solutions and advantages of this document, the technical solutions of the present application will be clearly and completely described below with reference to specific embodiments of the present application and corresponding drawings. It will be apparent that the embodiments described are only some, but not all, of the embodiments of this document. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are intended to be within the scope of this document.

Referring to fig. 1, an embodiment of the present application provides an identification method, including:

s102, acquiring an image sequence presented with the target object.

In this embodiment, the target object is any pedestrian represented in the image sequence.

In one possible way, the present embodiment extracts the image sequence from a compliant surveillance video.

In particular, a moving object in each image of the sequence of images may be identified based on the running object detector; the same moving object in each image is then identified based on existing motion feature prediction algorithms.

Based on the identification mode of the moving objects, the system can realize the respective locking of a plurality of moving objects in the image sequence, and the problem of capturing confusion is avoided.

S104, evaluating the images in the image sequence by taking the characteristic presentation quality of the target object as a standard.

In this embodiment, the feature presentation quality may include, but is not limited to: picture brightness, picture cut-off ratio, picture shielding degree, picture angle and the like of the target object.

Correspondingly, for each image in the image sequence, determining the index evaluation of each index in the characteristic presentation quality of the target object, and carrying out weighted summation on the index evaluation of all indexes to obtain the evaluation of the image.

S106, selecting a first preset number of representative images from the image sequence according to the evaluation result.

It should be understood that the quality of the feature presentation will directly lead to the accuracy of the base and the number of base obtained by comparison of the subsequent warehouse entry searches.

Therefore, according to the evaluation result, the embodiment can select the first preset number (not specifically limited herein) of representative images from the image sequence according to the order of the feature presentation quality from high to low, so as to be used for subsequent warehouse-in retrieval comparison.

S108, importing characteristic information (visual characteristic information, motion characteristic information and the like) of the target object in the representative image into a database for comparison to determine identity information of the target object; the database stores the corresponding relation between the characteristic information of the database and the identity information.

According to the embodiment, on the premise of compliance, the resident image and the identity information are put in storage in advance. The feature information of the base is the feature information extracted from the resident image, such as the visual feature information and the motion feature information.

Specifically, the present embodiment is divided into two stages to determine the identity information of the target object

One stage is to perform coarse screening from all the bottom libraries (bottom library feature information) of the database by adopting a low-computational-effort-cost vector-based similarity algorithm to determine a new bottom library similar to the feature information of the target object.

And in the other stage, a re-ranking algorithm with high calculation power overhead is adopted, the similarity between the characteristic information of the target object in the representative image and the characteristic information of each base in the new base is accurately calculated, and the identity information of the target object is determined according to the sequence from high to low of the similarity.

The process of determining identity information is described below in connection with one possible way.

1. Coarse screening stage

In this embodiment, first, based on a vector similarity algorithm, a second preset number of base feature information (text is not specifically limited) similar to feature information of the target object in the representative image may be determined from the database, so as to be used as first primary base feature information.

It should be understood that, because the similarity determined by the vector-based similarity algorithm is low, the first initially selected base feature information may not cover the target object, and therefore, a third preset number of other base feature information (text is not specifically limited) similar to the first initially selected base feature information may be determined from the database by using the vector-based similarity algorithm to serve as the second initially selected base feature information. And finding out the characteristic information of the second primary base which is highly similar to the characteristic information of the first primary base from the database.

And then determining the check base characteristic information of a fourth preset number (text is not specifically limited) according to the range of the first primary base characteristic information and the second primary base characteristic information. For example, determining an intersection between the first primary selected base characteristic information and the second primary selected base characteristic information, and determining the base characteristic information in the intersection as the check base characteristic information; for another example, a union between the first primary base feature information and the second primary base feature information is determined, and the pooled base feature information is determined as the check base feature information.

In practical application, the intersection mode or the union mode can be determined according to the number of the feature information of the first primary selection base. For example, when the number of the feature information of the first primary base is larger (reaching the preset threshold), an intersection mode may be adopted, and finally, the feature information of the check base is obtained and is a subset of the feature information of all the first primary base. For example, when the number of the feature information of the first primary base is larger (reaching the preset threshold), an intersection mode may be adopted, and finally, the feature information of the check base is obtained and is a subset of the feature information of all the first primary base. For another example, when the number of the feature information of the first primary base is small (the preset threshold is not reached), a union mode may be adopted, and finally, the feature information of the check base is obtained, which is based on all the feature information of the first primary base, and the feature information of the second primary base is supplemented.

The feature information of the check base obtained in this embodiment is a new base obtained after coarse screening.

2. Re-ranking algorithm stage

In this embodiment, a re-ranking algorithm is used to calculate the similarity between the feature information of the target object in the representative image and each of the feature information of the check base, and then at least one feature information of the target base is determined from the fourth preset number of feature information of the check base according to the order from high to low.

And then, determining the identity information of the target object in the identity information range corresponding to the characteristic information of the target base.

The method of the embodiment can be applied to identity recognition based on a cross-mirror tracking technology. Specifically, when an image sequence presenting a target object is obtained, evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard, so that a representative image with higher characteristic presentation quality is selected in the image sequence according to an evaluation result; and then, the characteristic information of the target object in the representative image is imported into a database to be searched and compared with the characteristic information of the base, and the base obtained by searching and comparing is more accurate due to higher characteristic presentation quality of the representative image, and finally, the relevant calculation of the identity recognition is carried out in the base with a small range, so that the efficiency of the identity recognition can be greatly improved.

The method of the embodiments is described below in connection with specific implementations.

The whole process of the method of the embodiment can be divided into four steps: video acquisition and processing, pedestrian detection and tracking, human body characteristic information presentation quality screening and warehousing and human body retrieval comparison.

1. Video acquisition and processing

The video acquisition is to acquire video files through video acquisition tools such as a monitoring camera, a network camera, a historical video file and the like.

The frame extraction means is to obtain an image at a certain frame rate interval (for example, one frame per second) in a section of video, and common frame extraction means include opencv frame extraction, video encoding and decoding, frame extraction software and the like.

2. Pedestrian detection and tracking

The first step of cross-mirror tracking is to detect the pedestrian position of an image seed, the application adopts a deep neural network yolov to realize human body detection, in order to avoid that the same target enters a search base for many times, a pedestrian tracking technology is needed to carry out primary screening, and the tracking algorithm comprises the following steps:

1) Given an original frame of video: extracting frames from each video file obtained in the step one to obtain each frame of image in the video file, namely an original frame;

2) Running the object detector to identify each frame of image to obtain a bounding box of the object;

3) For each detected object, different features, typically visual and motion features, are computed;

4) According to the vision and the motion characteristics, firstly, selecting a boundary frame of which the front frame and the rear frame are possibly the same object through a motion characteristic prediction algorithm (such as an IOU matching algorithm), and then calculating the probability that two objects belong to the same object through a similarity calculation algorithm (such as a characteristic matching algorithm);

5) Each object is marked with a digital ID.

3. Human body characteristic information presentation quality screening and warehousing

After the detection and tracking of the pedestrians are performed, each pedestrian can obtain a group of image sequences, a human body image in the image sequences is calculated by using a human body characteristic information presentation quality judgment algorithm according to the generated image sequences, the characteristic information presentation quality scores of the human body are given, if the characteristic information presentation quality scores of the human body in the image sequences are lower than a threshold value (for example, 0.5), the group of images are abandoned, otherwise, the human body with the highest quality score is selected for warehousing.

The human body characteristic information presentation quality judgment algorithm is mainly comprehensively evaluated through the image brightness, the image size, the image cut-off ratio, the image shielding degree and the image angle of the human body.

1) Human body picture brightness calculation

The brightness of the picture is the most intuitive feeling of an image to people, and various methods for calculating the brightness of the picture can be used for converting an RGB color space into a YUV color space, wherein the average value of Y can represent the brightness of the image, and the calculation formula is as follows:

Y＝mean(0.299R+0.587G+0.114B)

after calculating the brightness of the image, setting an image brightness scoring rule according to the visual experience of a person, wherein the calculation formula is as follows:

2) Human body picture size calculation

The picture size refers to the size of a human frame detected by a pedestrian, namely, the visual reflection of the definition of the human body, and the calculation formula is as follows:

s＝w*h

Wherein w is the width of the human body frame, and h is the height of the human body frame.

The calculation formula of the picture size score is as follows:

3) Human picture cut-off ratio calculation

The picture cut-off ratio refers to the aspect ratio of the human body image, namely the direct reflection of the human body integrity, and the mania-height ratio of the human body image is 1:5 to 1:4 is the best, and the calculation formula is as follows:

ARt＝w/h

the scoring calculation formula of the picture cut-off ratio is as follows:

4) Judgment of the degree of shielding of human body

The picture shielding degree refers to that a human body is shielded by other people or objects, namely, the accuracy of the human body is directly reflected, and the picture shielding degree is classified into three types by adopting a deep neural network: no shielding, slight shielding and serious shielding;

the scoring calculation formula of the picture shielding degree is as follows:

5) Human body picture angle judgment

The picture angle judgment refers to the direction of a human body in an image, and the method adopts a deep neural network to classify the direction of the human body into six types: front, back, front to right, front to left, back to right, and stool to left;

The calculation formula of the picture angle judgment score is as follows:

After the quality judgment scores of the characteristic information presentation of the five human bodies are obtained, the specification of the final score rule of the characteristic information presentation quality of the human bodies can be carried out according to the actual scene requirements, for example, the influence of human body truncation on a service system is large, and the human body truncation can be given higher weight, and the calculation formula is as follows:

score＝0.1*score_light+0.1*score_area+0.2*score_occ

+0.2*socre_angle+0.4*score_Art

4. Human body retrieval and comparison

The human body retrieval comparison is to extract the feature information of the query image (representative image) and the retrieval base image through the deep neural network, output the feature information as a feature vector with fixed dimension, calculate the similarity distance between the features of the query image and the base image according to the similarity algorithm of the vector, such as cosine similarity, and select a similar target topK (such as k=5) as a retrieval result according to the sequence of the similarity distances. The corresponding flow is shown in fig. 3, comprising:

Feature distances between the query image (representative image) and the base image are calculated through a similarity algorithm of vectors. The similarity calculation formula is as follows:

Wherein A, B denotes feature vectors of the query image and the base image, the sum of a and B represents the length of the feature vector.

After the feature similarity is calculated, a feature distance matrix A can be obtained, wherein the dimension of the matrix A is 1 x m, m is the number of the bottom library images, and the feature distance matrix A contains the feature information of the first primary selection bottom library.

Then, selecting the bottom library image of N before feature distance sorting as a new query image, and then respectively carrying out similarity calculation on N new query images and the bottom library image (in order to reduce post-processing time, the bottom library image comprises N new query images) through a vector similarity algorithm to obtain a feature distance matrix B; at this time, the dimension of the feature distance matrix B is n×m, which includes the second primary base feature information described above.

Then, splicing the characteristic distance matrix A and the characteristic distance matrix B to obtain a characteristic distance matrix C, wherein the dimension of the matrix C is (N+1) m; then, features of K before feature distance sequencing are intercepted from a feature distance matrix C to serve as a retrieval base of a coarse screen, and a feature matrix D is obtained, wherein the dimension of the feature matrix D is (N+1) K.

Next, the base of the feature matrix D is further screened, which can be classified into two cases:

Case one

And carrying out intersection operation on the characteristic matrix D in a first dimension, wherein the corresponding formula is as follows:

wherein E is equivalent to the intersection between the first primary base characteristic information and the second primary base characteristic information, and comprises the check base characteristic information, D is a characteristic matrix obtained in the first stage, i represents an index of D in a first dimension, N represents the first dimension of D, and D (i: i) represents taking all elements of an ith row.

It should be appreciated that intersection processing can simplify the base to the common result of all query pictures, and can effectively construct a high-quality base.

Case two

And carrying out union operation on the characteristic matrix in a first dimension, wherein the formula is as follows:

in the formula, E is equivalent to a union set between the first primary selection base feature information and the second primary selection base feature information, including the aforementioned check base feature information, and the meaning of other symbols is the same as that of the intersection operation formula, which is not described herein again.

It is understood that the union processing can save the base graphs inquired by different inquiry graphs as much as possible on the premise of removing repeated pictures, and the operation can obtain rich and stronger base;

After the E is obtained, calculating the feature distance between the feature information of the original query image and the feature information of the check base in the E through a re-ranking algorithm, sorting according to the similarity represented by the feature distance to obtain a final feature distance matrix, and screening the feature information of the check base with the similarity reaching a certain standard from the final feature distance matrix to define the feature information of the target base.

And finally, finding out the identity information of the human body in the query image in the identity information range corresponding to the characteristic information of the target base.

In addition, referring to fig. 4, another embodiment of the present application further provides an identification device 400, including:

The acquisition module 310 acquires a sequence of images that are presented with a target object.

And the evaluation module 320 evaluates the images in the image sequence by taking the characteristic presentation quality of the target object as a standard.

And a selecting module 330, configured to select a first preset number of representative images from the image sequence according to the evaluation result.

The comparison module 340 imports the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

The device of the embodiment can be applied to identity recognition based on a cross-mirror tracking technology. Specifically, when an image sequence presenting a target object is obtained, evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard, so that a representative image with higher characteristic presentation quality is selected in the image sequence according to an evaluation result; and then, the characteristic information of the target object in the representative image is imported into a database to be searched and compared with the characteristic information of the base, and the base obtained by searching and comparing is more accurate due to higher characteristic presentation quality of the representative image, and finally, the relevant calculation of the identity recognition is carried out in the base with a small range, so that the efficiency of the identity recognition can be greatly improved.

Optionally, the feature presentation quality includes at least one of the following indicators:

The picture brightness of the target object;

The picture size of the target object;

The picture cut-off ratio of the target object;

The picture shielding degree of the target object;

and the picture angle of the target object.

Optionally, the evaluating module 320 evaluates the images in the image sequence based on the characteristic presentation quality of the target object, including: and determining index evaluation of each index in the characteristic presentation quality of the target object for each image in the image sequence, and carrying out weighted summation on the index evaluation of all indexes to obtain the evaluation of the image.

Optionally, the comparing module 340 imports the feature information of the target object in the representative image into a database for comparison to determine the identity information of the target object, including: determining a second preset number of base characteristic information similar to the characteristic information of the target object in the representative image from the database by using a vector-based similarity algorithm, and taking the second preset number of base characteristic information as first primary base characteristic information; determining a third preset number of other base characteristic information similar to the first primary base characteristic information from the database by using a vector-based similarity algorithm, and taking the third preset number of other base characteristic information as second primary base characteristic information; determining a fourth preset number of check base characteristic information in the range of the first primary base characteristic information and the second primary base characteristic information; determining target base characteristic information similar to the characteristic information of the target object in the representative image from fourth preset number of check base characteristic information by using a re-ranking algorithm; and determining the identity information of the target object in the identity information range corresponding to the characteristic information of the target base.

Optionally, the comparing module 340 determines a fourth preset number of check base feature information within the range of the first primary base feature information and the second primary base feature information, including: determining an intersection between the first primary selection base characteristic information and the second primary selection base characteristic information, and determining the base characteristic information in the intersection as check base characteristic information; or determining a union between the first primary base characteristic information and the second primary base characteristic information, and determining the library characteristic information in the union as the check base characteristic information.

Optionally, the apparatus of this embodiment further includes:

The marking module is used for identifying the moving object in each image of the image sequence based on the running object detector and identifying the same moving object in each image based on a motion feature prediction algorithm before the evaluating module 320 evaluates the images in the image sequence by taking the characteristic presentation quality of the target object as a standard; wherein the target object belongs to an identified moving object.

Optionally, the characteristic information of the target object in the representative image includes visual characteristic information and motion characteristic information.

It should be noted that, the identity recognition device of the present embodiment may be used as an execution subject of the method shown in fig. 1, so that the steps and functions of the method shown in fig. 1 can be implemented.

Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application. Referring to fig. 4, at the hardware level, the electronic device includes a processor, and optionally an internal bus, a network interface, and a memory. The Memory may include a Memory, such as a Random-Access Memory (RAM), and may further include a non-volatile Memory (non-volatile Memory), such as at least 1 disk Memory. Of course, the electronic device may also include hardware required for other services.

The processor, network interface, and memory may be interconnected by an internal bus, which may be an ISA (Industry Standard Architecture ) bus, a PCI (PERIPHERAL COMPONENT INTERCONNECT, peripheral component interconnect standard) bus, or EISA (Extended Industry Standard Architecture ) bus, among others. The buses may be classified as address buses, data buses, control buses, etc. For ease of illustration, only one bi-directional arrow is shown in FIG. 4, but not only one bus or type of bus.

And the memory is used for storing programs. In particular, the program may include program code including computer-operating instructions. The memory may include memory and non-volatile storage and provide instructions and data to the processor.

The processor reads the corresponding computer program from the nonvolatile memory to the memory and then runs the computer program to form the identification device on a logic level. The processor is used for executing the programs stored in the memory and is specifically used for executing the following operations:

A sequence of images presenting a target object is acquired.

And evaluating the images in the image sequence by taking the characteristic presentation quality of the target object as a standard.

And selecting a first preset number of representative images from the image sequence according to the evaluation result.

Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

The method disclosed in the embodiment of fig. 1 of the present application can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or by instructions in the form of software. The processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; but may also be a digital signal Processor (DIGITAL SIGNAL Processor, DSP), application SPECIFIC INTEGRATED Circuit (ASIC), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware components. The disclosed methods, steps, and logic diagrams in one or more embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with one or more embodiments of the application may be embodied directly in hardware, in a decoded processor, or in a combination of hardware and software modules in a decoded processor. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The storage medium is located in a memory, and the processor reads the information in the memory and, in combination with its hardware, performs the steps of the above method.

Of course, other implementations, such as a logic device or a combination of hardware and software, are not excluded from the electronic device of the present application, that is, the execution subject of the following processing flows is not limited to each logic unit, but may be hardware or a logic device.

The embodiments of the present application also propose a computer program product comprising a computer-readable storage medium storing a computer program, the computer program being operable to cause a computer to:

A sequence of images presenting a target object is acquired.

And evaluating the images in the image sequence by taking the characteristic presentation quality of the target object as a standard.

And selecting a first preset number of representative images from the image sequence according to the evaluation result.

Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

In summary, the foregoing description is only of the preferred embodiments of the present application, and is not intended to limit the scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of one or more embodiments of the present application should be included in the protection scope of one or more embodiments of the present application.

The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. One typical implementation is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.

The embodiments of the present application are described in a progressive manner, and the same and similar parts of the embodiments are all referred to each other, and each embodiment is mainly described in the differences from the other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.

Claims (10)

1. An identification method, comprising:

acquiring an image sequence presenting a target object;

Evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard;

Selecting a first preset number of representative images from the image sequence according to the evaluation result;

Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

2. The method according to claim 1,

The feature presentation quality includes at least one of the following indicators:

The picture brightness of the target object;

The picture size of the target object;

The picture cut-off ratio of the target object;

The picture shielding degree of the target object;

and the picture angle of the target object.

3. The method according to claim 2,

Evaluating the images in the image sequence by taking the characteristic presentation quality of the target object as a standard, wherein the evaluating comprises the following steps:

And determining index evaluation of each index in the characteristic presentation quality of the target object for each image in the image sequence, and carrying out weighted summation on the index evaluation of all indexes to obtain the evaluation of the image.

4. The method according to claim 1,

Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object, wherein the method comprises the following steps:

determining a second preset number of base characteristic information similar to the characteristic information of the target object in the representative image from the database by using a vector-based similarity algorithm, and taking the second preset number of base characteristic information as first primary base characteristic information;

Determining a third preset number of other base characteristic information similar to the first primary base characteristic information from the database by using a vector-based similarity algorithm, and taking the third preset number of other base characteristic information as second primary base characteristic information;

Determining a fourth preset number of check base characteristic information in the range of the first primary base characteristic information and the second primary base characteristic information;

Determining target base characteristic information similar to the characteristic information of the target object in the representative image from fourth preset number of check base characteristic information by using a re-ranking algorithm;

And determining the identity information of the target object in the identity information range corresponding to the characteristic information of the target base.

5. The method according to claim 4, wherein the method comprises,

Determining a fourth preset number of check base feature information within the range of the first primary base feature information and the second primary base feature information, including:

Determining an intersection between the first primary selection base characteristic information and the second primary selection base characteristic information, and determining the base characteristic information in the intersection as check base characteristic information;

Or alternatively

And determining a union between the first primary selection base characteristic information and the second primary selection base characteristic information, and determining the base characteristic information in the union as check base characteristic information.

6. The method according to claim 1 to 5,

Before evaluating the images in the image sequence based on the characteristic presentation quality of the target object, the method further comprises:

identifying a moving object in each image of the sequence of images based on a running object detector;

Identifying the same moving object in each image based on a motion characteristic prediction algorithm; wherein the target object belongs to an identified moving object.

7. The method according to claim 1 to 5,

The characteristic information of the target object in the representative image includes visual characteristic information and motion characteristic information.

8. An identification device, comprising:

the acquisition module acquires an image sequence presenting a target object;

The evaluation module is used for evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard;

The selecting module is used for selecting a first preset number of representative images from the image sequence according to the evaluation result;

The comparison module is used for importing the characteristic information of the target object in the representative image into a database for comparison so as to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

9. An electronic device, comprising: a processor; and a memory arranged to store computer executable instructions that, when executed, cause the processor to:

acquiring an image sequence presenting a target object;

Evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard;

Selecting a first preset number of representative images from the image sequence according to the evaluation result;

Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.

10. A computer program product comprising a computer readable storage medium storing a computer program operable to cause a computer to:

acquiring an image sequence presenting a target object;

Evaluating images in the image sequence by taking the characteristic presentation quality of the target object as a standard;

Selecting a first preset number of representative images from the image sequence according to the evaluation result;

Importing the characteristic information of the target object in the representative image into a database for comparison to determine the identity information of the target object; the database stores the corresponding relation between the characteristic information and the identity information of the database.