A multi-algorithm fusion image comparison method and system based on an ensemble algorithm

Abstract

The application provides a multi-algorithm fusion image comparison method based on an integrated algorithm, which comprises the following steps: S1, acquiring a picture set and respectively dividing pictures of the same object into static sets and dynamic sets, and giving the pictures of the same object an associated actorId; S2, respectively extracting picture feature vectors of the static sets and the dynamic sets by using multiple algorithms to obtain basic features and derived features; S3, constructing samples based on the obtained basic features and derived features and the associated actorId; S4, substituting the samples into a preset integrated algorithm to train, evaluate and save an optimal model; and S5, inputting the obtained basic features and derived features of the pictures into the saved optimal model to obtain a prediction result. By using the integrated algorithm to learn the characteristics among multiple algorithms, the influence of the subjective experience of the weighting strategy on the overall situation can be compensated for the shortcomings of a single algorithm, the model generalization capability is improved, the image comparison accuracy and recall rate are improved, and reliable comparison services are provided for image clustering.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of security and defense, and particularly relates to a multi-algorithm fusion image comparison method and system based on an integrated algorithm. BACKGROUND

[0002] In the field of security and defense, with the development of image recognition technology, although the image recognition technology has high accuracy, it is still affected by various environmental conditions, and each recognition algorithm will cause certain misidentification and missed identification, and has weak generalization ability.

[0003] Further, in order to improve the system accuracy, users usually select multiple different recognition algorithms for complementary advantages, and determine the weighting proportion through subjective experience weighting fusion or through statistical methods, so as to improve the system image recognition rate. Although the image recognition rate is improved, the improvement space is limited, and the generalization ability is limited. Only the advantages and disadvantages of the algorithm are considered to add different proportion weights, and the potential mutual information between algorithms cannot be deeply mined.

[0004] Therefore, it is of great significance to propose a multi-algorithm fusion image comparison method and system based on an integrated algorithm. SUMMARY

[0005] In order to solve the problems of low accuracy, low recall rate and low generalization ability in the existing security and defense field using single algorithm identification comparison, the fusion comparison of expert weighting or statistical weighting method cannot deeply mine the internal potential information based on experience or simple statistical method, has low generalization ability, and has insufficient robustness, etc. The present application provides a multi-algorithm fusion image comparison method and system based on an integrated algorithm to solve the above technical defects.

[0006] In a first aspect, the present application provides a multi-algorithm fusion image comparison method based on an integrated algorithm, which comprises the following steps:

[0007] S1, acquiring a picture set and dividing pictures of the same object in the picture set into a static set and a dynamic set, and assigning the pictures of the same object with an associated actorId;

[0008] S2, using multiple algorithms to extract picture feature vectors of the static set and the dynamic set respectively to obtain basic features and derived features;

[0009] S3, constructing a sample based on the obtained basic features, derived features and associated actorId;

[0010] S4, substituting the sample into a preset integrated algorithm for further training, evaluation and saving of the best model; and

[0011] S5, input the base features and the derived features of the obtained pictures into the saved best model to obtain a prediction result.

[0012] Preferably, in S1, further comprising: using the static set as a comparison library and the dynamic set as a search picture to obtain a comparison score by performing 1:N similarity calculation between the search picture and the pictures in the static set.

[0013] Preferably, in S2, further comprising: performing feature vector similarity calculation between each two pictures in the static set and the dynamic set, wherein the score of each algorithm is used as the base feature, and the derived feature is generated based on the effect comparison between algorithms.

[0014] Further preferably, in S3, further comprising: dividing each sample into two parts of score feature and label, wherein the label of the sample is determined by the associated actorId, if the associated actorId of the search picture and the picture in the comparison static set is consistent, the label is assigned as 1, otherwise the label is assigned as 0.

[0015] Further preferably, the specific generation of the derived feature comprises:

[0016] S21, obtaining the score and the label of each algorithm, and splicing the search picture id and the comparison result picture id to obtain the sample index_id, merging the score results of different algorithms with the same search picture and comparison result picture to obtain an initial sample in the form of <_id, alg_a, alg_b, …, algN, label>, wherein algN is the score similarity value of each algorithm, and label is the label;

[0017] S22, respectively determining the relative importance of each two different algorithms, and defining the similarity threshold <T1, T2, …, TN> of each algorithm.

[0018] Preferably, in S4, specifically comprising:

[0019] The constructed samples are respectively fed into different integrated algorithms such as random forest, LGB and GBDT, the best model parameters under each integrated algorithm are confirmed according to cross-validation, the best algorithm under each integrated algorithm is saved, and finally the best integrated algorithm is selected as the final model for image comparison by evaluating the effect of the test set.

[0020] Further preferably, in S5, specifically comprising:

[0021] According to the base features obtained by multiple algorithms and the derived features, the best integrated algorithm model is input to obtain the final fusion score result, and if the result exceeds the first probability threshold of the model, it is considered as the same object.

[0022] In a second aspect, the embodiments of the present application further provide a multi-algorithm fusion image comparison system based on an ensemble algorithm, comprising:

[0023] An acquisition module is configured to acquire a picture set.

[0024] A division module is configured to divide pictures of the same object in the picture set into a static set and a dynamic set.

[0025] An association module is configured to assign an association actorId to the pictures of the same object, and determine a label label of a sample based on the association actorId.

[0026] An extraction module is configured to extract picture feature vectors of the static set and the dynamic set by using multiple algorithms, to obtain basic features and derived features.

[0027] A sample construction module is configured to construct a sample based on the obtained basic features, derived features and association actorId.

[0028] A training module is configured to further train, evaluate and save an optimal model by substituting the sample into a preset ensemble algorithm.

[0029] A prediction module is configured to input the basic features and derived features of an obtained picture into the saved optimal model, to obtain a prediction result.

[0030] In a third aspect, the embodiments of the present application provide an electronic device, comprising: one or more processors; a storage device configured to store one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any of the implementation manners of the first aspect.

[0031] In a fourth aspect, the embodiments of the present application provide a computer readable storage medium, which stores a computer program, when the computer program is executed by a processor, the method described in any of the implementation manners of the first aspect is implemented.

[0032] Compared with the prior art, the beneficial results of the present application are as follows:

[0033] (1) By using the ensemble algorithm to learn the characteristics between multiple algorithms, the shortcomings of single algorithm and the subjective experience of weighting strategy can be compensated, the mutual influence between multiple algorithm comparison parts is deeply discovered and complemented, the potential mutual relationship between algorithms is mined, the model generalization ability is improved, and then the image comparison accuracy and recall rate are improved, and reliable comparison services are provided for image clustering.

[0034] (2) The technical scheme of the present application defines a derived feature generated by comparing the performance difference of different algorithms, and combines the comparison score obtained by multiple algorithms as a sample feature, uses an integrated algorithm to learn the characteristics between multiple algorithms to train a model, and then gives a fusion score of image comparison to judge the similarity between images, thereby improving the precision and recall rate of comparison. BRIEF DESCRIPTION OF DRAWINGS

[0035] The accompanying drawings are included to provide a further understanding of embodiments and are incorporated in and constitute a part of this specification. The drawings illustrate embodiments and, together with the description, serve to explain the principles of the present application. Other embodiments and many of the intended advantages of the present application will be readily appreciated as the same becomes better understood by reference to the following detailed description when considered in connection with the accompanying drawings. The elements of the drawings are not necessarily to scale relative to each other. Like reference numerals designate corresponding similar parts.

[0036] Figure 1 is an exemplary device architecture diagram in which an embodiment of the present application can be applied;

[0037] Figure 2 is a flowchart of an integrated algorithm-based multi-algorithm fusion image comparison method of an embodiment of the present application;

[0038] Figure 3 is a main flowchart of an integrated algorithm-based multi-algorithm fusion image comparison method of an embodiment of the present application;

[0039] Figure 4 is a schematic diagram of picture set division in an integrated algorithm-based multi-algorithm fusion image comparison method of an embodiment of the present application;

[0040] Figure 5 is a structural schematic diagram of an integrated algorithm-based multi-algorithm fusion image comparison system of an embodiment of the present application;

[0041] Figure 6 is a structural schematic diagram of a computer device of an electronic device suitable for implementing an embodiment of the present application. DETAILED DESCRIPTION

[0042] In the following detailed description, reference is made to the accompanying drawings, which form a part hereof, and in which are shown by way of illustration illustrative embodiments in which the application can be practiced. For purposes of explanation and illustration, directional terms are used with reference to the orientation of the described figures. However, it is to be understood that the embodiments can be practiced in other orientations than those described and shown. The directional terms used herein refer only to the direction of the components illustrated in the figures and do not limit the scope of the application to only such oriented structures and acts. Other embodiments can be used and / or other changes can be made without departing from the scope of the application. The following detailed description is, therefore, not to be taken in a limiting sense, as the scope of the application is defined by the appended claims.

[0043] It should be understood that Figure 1 The number of terminal devices, networks and servers in the system is merely illustrative. Depending on the implementation needs, there can be any number of terminal devices, networks and servers.

[0044] Figure 1 An exemplary system architecture 100 for a method for processing information or an apparatus for processing information to which embodiments of the present application can be applied is shown.

[0045] As Figure 1 The system architecture 100 can include terminal devices 101, 102, 103, a network 104 and a server 105, as shown. The network 104 is a medium for providing a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 can include various connection types, such as wired, wireless communication links or fiber optic cables, etc.

[0046] A user can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications can be installed on the terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0047] The terminal devices 101, 102, 103 can be various electronic devices with communication functions, including but not limited to smart phones, tablet computers, laptop computers and desktop computers, etc.

[0048] The server 105 can be a server that provides various services, such as a background information processing server that processes verification request information sent by the terminal devices 101, 102, 103. The background information processing server can analyze and process the received verification request information, and obtain a processing result (such as verification success information for indicating that the verification request is a legal request).

[0049] It should be noted that the method for processing information provided by the embodiment of the present application is generally executed by the server 105, and accordingly, the device for processing information is generally arranged in the server 105. In addition, the method for sending information provided by the embodiment of the present application is generally executed by the terminal device 101, 102 or 103, and accordingly, the device for sending information is generally arranged in the terminal device 101, 102 or 103.

[0050] It should be noted that the server can be hardware or software. When the server is hardware, it can be implemented as a distributed server cluster composed of multiple servers or as a single server. When the server is software, it can be implemented as multiple software or software modules (for example, to provide distributed services) or as a single software or multiple software modules, which are not specifically limited here.

[0051] In the field of security, with the development of image recognition technology, although the image recognition technology has high accuracy, it is still affected by various environmental conditions, and each recognition algorithm will cause certain misidentification and missed identification, and has weak generalization ability. Further, in order to improve the system accuracy, the user usually selects multiple different recognition algorithms to complement each other, and determines the weighting proportion through subjective experience weighting fusion or through statistical method, so as to improve the system image recognition rate. Although the image recognition rate is improved, the improvement space is limited, and the generalization ability is limited, only the advantages and disadvantages of the algorithm itself are considered to add different proportion weights, and the potential mutual information between algorithms cannot be deeply mined.

[0052] The existing related technologies have the following disadvantages:

[0053] 1. Single algorithm recognition comparison, relatively low accuracy and recall rate, and low generalization ability;

[0054] 2. Fusion comparison by expert weighting or statistical weighting method, based on experience or simple statistical method, which cannot deeply mine internal potential information, has low generalization ability and insufficient robustness.

[0055] In order to overcome these defects and problems, the present application provides a multi-algorithm fusion image comparison method and system based on an integrated algorithm, which uses an integrated algorithm to learn the characteristics between multiple algorithms, has strong anti-interference ability, strong model generalization ability, can balance the error of unbalanced data set, can detect the mutual influence between multiple algorithm comparison parts and complement each other, and greatly improves the accuracy and recall rate.

[0056] In a first aspect, Figure 2 An embodiment of the present application is shown to disclose a flowchart of a multi-algorithm fusion image comparison method based on an integrated algorithm, as shown in Figure 2 and Figure 3 The method comprises the following steps:

[0057] S1, obtaining a picture set and dividing pictures of the same object in the picture set into a static set and a dynamic set respectively, and assigning the pictures of the same object with an associated actorId;

[0058] Specifically, in the embodiment, in S1, further comprising: taking the static set as a comparison library and the dynamic set as a search picture, and using the search picture to perform 1:N similarity calculation with the pictures in the static set to obtain a comparison score, in this embodiment, the comparison score refers to the similarity value between two pictures, ranging from 0 to 1. As shown in the following table. Figure 4

[0059] S2, extracting picture feature vectors of the static set and the dynamic set respectively by using multiple algorithms to obtain basic features and derived features;

[0060] Specifically, in S2, further comprising: performing feature vector similarity calculation between two pictures in the static set and the dynamic set, wherein the score of each algorithm is taken as the basic feature, and the derived feature is generated based on the effect comparison between algorithms. In this embodiment, the score of each algorithm refers to the similarity value between two pictures compared by each algorithm, ranging from 0 to 1.

[0061] S3, constructing samples based on the obtained basic features, derived features and associated actorId;

[0062] Further, in S3, further comprising: each of the samples is divided into two parts of score feature and label, wherein the label of the sample is determined by the associated actorId, if the associated actorId of the search picture and the compared picture in the static set is consistent, the label 1 is assigned, otherwise the label 0 is assigned.

[0063] Specifically, the specific generation of the derived feature includes:

[0064] S21, obtaining the score and label of each algorithm, and splicing the search picture id and the comparison result picture id as the sample index_id, merging the score results of different algorithms with the same search picture and comparison result picture to obtain an initial sample in the form of <id, alg_a, alg_b, …, algN, label>, wherein algN is the score similarity value of each algorithm, and label is the label;

[0065] S22, respectively judging the relative importance of two different algorithms, and defining the similarity threshold <T1, T2, …, TN> of each algorithm.

[0066] ​In the embodiment, taking the importance of algorithm a relative to algorithm b as an example, then there are steps:

[0067] Step 1. Filter the samples that satisfy the threshold of algorithm a and algorithm b;

[0068] Step 2. Calculate the number of samples TP1 that satisfy the threshold of algorithm a and are labeled as 1, the number of samples FP1 that satisfy the threshold of algorithm a but are labeled as 0, the number of samples FN1 that do not satisfy the threshold of algorithm a but are labeled as 1, and the number of samples TN1 that do not satisfy the threshold of algorithm a and are labeled as 0, respectively, then there are precision pre_aob = TP1 / (TP1+FP1), recall rec_aob = TP1 / (TP1+FN1), and the index value of the binary classification model F1_aob = 2*pre_aob*rec_aob / (pre_aob+rec_aob). Similarly, the index value of the binary classification model F1_boa = 2*pre_boa*rec_boa / (pre_boa+rec_boa) of algorithm b can be obtained.

[0069] Step 3. Determine whether F1_aob is greater than F1_boa, if greater than or equal to, then generate the derived feature alg_ab = (alg_a+k)*log2(1+alg_b+k), if less than, then generate the derived feature alg_ab = (alg_b+k)*log2(1+alg_a+k), where k is a value that has not been avoided to avoid the case that alg_a is 0.

[0070] Step 4. Similarly, the corresponding alg_NM series of derived feature calculation formulas are obtained by pairwise determination between other algorithms.

[0071] S4, substituting the sample into the preset integrated algorithm for further training, evaluation and saving the best model; and

[0072] Specifically, in S4, specifically includes: feeding the constructed sample into different integrated algorithms such as random forest, LGB, GBDT and other different integrated algorithms, confirming the best model parameters under each integrated algorithm according to cross-validation, saving the best algorithm under each integrated algorithm, and finally selecting the best integrated algorithm as the final model for image comparison through the evaluation effect of the test set.

[0073] S5, inputting the obtained basic features and derived features of the picture into the saved best model to obtain a prediction result.

[0074] Further, in S5, specifically comprising: according to the plurality of algorithms, obtaining the basic features, and inputting the derived features obtained into the best integrated algorithm model to obtain a final fusion score result, and considering that the same object is more than the first probability threshold of the model. In this embodiment, the first probability threshold of the model refers to the first threshold of the fusion similarity score, and the threshold range is [0, 1].

[0075] In a second aspect, the embodiments of the present application also provide a multi-algorithm fusion image comparison system based on an integrated algorithm, as shown in the accompanying drawings, comprising: Figure 5

[0076] The acquisition module 51 is configured to acquire a picture set.

[0077] The division module 52 is configured to divide the pictures of the same object in the picture set into a static set and a dynamic set.

[0078] The association module 53 is configured to assign an associated actorId to the pictures of the same object, and determine the label of the sample by the associated actorId.

[0079] The extraction module 54 is configured to extract the picture feature vectors of the static set and the dynamic set by using a plurality of algorithms respectively, so as to obtain the basic features and the derived features.

[0080] The sample construction module 55 is configured to construct a sample based on the obtained basic features, derived features and associated actorId.

[0081] The training module 56 is configured to further train, evaluate and save the best model by substituting the sample into a preset integrated algorithm.

[0082] The prediction module 57 is configured to input the basic features and derived features of the obtained picture into the saved best model to obtain a prediction result.

[0083] The technical scheme of the present application defines a derived feature generated by comparing the performance differences of different algorithms, and combines the comparison scores obtained by a plurality of algorithms as sample features, uses an integrated algorithm to learn the characteristics between the plurality of algorithms to train a model, and then gives a fusion score of image comparison, so as to judge the similarity between images, and thus improve the precision and recall rate of comparison.

[0084] By using the integrated algorithm to learn the characteristics between the plurality of algorithms, the shortcomings of the single algorithm and the subjective experience of the weighting strategy can be compensated, the mutual influence between the comparison scores of the plurality of algorithms is deeply discovered and complemented, the potential mutual relationship between the algorithms is mined, the generalization ability of the model is improved, and thus the image comparison accuracy and recall rate are improved, and reliable comparison services are provided for image clustering. ​

[0085] Reference will now be made to the drawings Figure 6 which show an electronic device (e.g., a server or terminal device) suitable for use in implementing embodiments of the present application. Figure 1 A block diagram of a computer arrangement 600 of the electronic device shown is shown. Figure 6 The electronic device shown is merely an example and should not limit the function and scope of use of embodiments of the present application in any way.

[0086] As shown in Figure 6 , the computer arrangement 600 includes a central processing unit (CPU) 601 and a graphics processor (GPU) 602, which can perform various appropriate actions and processes in accordance with programs stored in a read only memory (ROM) 603 or programs loaded from a storage section 609 into a random access memory (RAM) 606. In the RAM 604, various programs and data required for the operation of the arrangement 600 are also stored. The CPU 601, the GPU 602, the ROM 603, and the RAM 604 are connected to each other through a bus 605. An input / output (I / O) interface 606 is also connected to the bus 605.

[0087] The following components are connected to the I / O interface 606: an input section 607 including a keyboard, a mouse, etc.; an output section 608 including a display such as a cathode ray tube (CRT) display, a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 609 including a hard disk, etc.; and a communication section 610 including a network interface card such as a LAN card, a modem, etc. The communication section 610 performs communication processing via a network such as the Internet. A drive 611 can also be connected to the I / O interface 606 as necessary. A removable media 612 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is attached to the drive 611 as necessary, so that a computer program read therefrom is installed into the storage section 609 as necessary.

[0088] In particular, in accordance with embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network by the communication section 610, and / or installed from the removable media 612. When the computer program is executed by the central processing unit (CPU) 601 and the graphics processor (GPU) 602, the above-described functions defined in the methods of the present application are performed.

[0089] It should be noted that the computer-readable medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the above. More specific examples of the computer-readable medium can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer-readable medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device. In the present application, the computer-readable signal medium can include a data signal that propagates in a baseband or as part of a carrier wave, in which the computer-readable program code is carried. Such a propagated data signal can take any of a variety of forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium that is not a computer-readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including, but not limited to, wireless, wireline, optical fiber, RF, etc., or any suitable combination of the above.

[0090] Computer program code for carrying out operations of the present application can be written in any combination of one or more programming languages, including an object-oriented programming language such as Java, Smalltalk, C++, or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0091] The computer readable medium can be a storage medium or memory medium such as, but not limited to, readable and / or writeable nonvolatile memory devices (e.g., EPROM, EEPROM, Flash memory, etc.); volatile storage devices (e.g., RAM such as SRAM, DRAM, zero bus, etc.); or any memory storage devices suitable for the storage and / or transmission of information.

[0092] The modules described in the embodiments of the present application can be implemented by software or by hardware. The modules described can be located in a processor.

[0093] As another aspect, the present application also provides a computer readable medium, which can be included in the electronic device described in the above embodiments, or can exist separately without being assembled into the electronic device. The computer readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the method steps described in the first aspect.

[0094] The above description is merely illustrative of the embodiments of the present application and the principles of the technology employed. It is understood that the scope of the present application is not limited to the specific combinations of technical features described above, but also covers other technical solutions formed by any combination of the technical features described above or equivalent features, without departing from the inventive concept. For example, the technical solutions formed by replacing the above features with technical features disclosed in the present application (but not limited to) having similar functions.

Claims

1. A multi-algorithm fusion image registration method based on integrated algorithm, characterized in that, The method comprises the following steps: S1, acquiring a picture set and dividing pictures of the same object in the picture set into a static set and a dynamic set respectively, and assigning the pictures of the same object with an associated actorId; using the search picture to perform 1:N similarity calculation with the pictures in the static set to obtain a comparison score; S2, extracting picture feature vectors of the static set and the dynamic set respectively using a preset algorithm to obtain basic features and derived features; performing feature vector similarity calculation between every two pictures in the static set and the dynamic set, wherein the scores of each algorithm are taken as the basic features, and the derived features are generated based on the effect comparison between algorithms; wherein the specific generation of the derived features comprises: S21, acquiring the scores and labels of each algorithm, and concatenating the search picture id and the comparison result picture id as a sample index_id, merging the score results of different algorithms with the same search picture and comparison result picture to obtain an initial sample in the form of <_id, alg_a, alg_b, …, algN, label>, wherein algN is the score similarity value of each algorithm, and label is the label; S22, respectively judging the relative importance of two different algorithms, and defining the similarity threshold values <T1, T2, …, TN> of each algorithm; S3, constructing samples based on the obtained basic features and derived features and the associated actorId; each sample is divided into two parts of score features and labels, wherein the label of the sample is determined by the associated actorId, if the associated actorId of the search picture and the pictures in the static set for comparison is consistent, then the label is assigned as 1, otherwise the label is assigned as 0; S4, substituting the samples into a preset integrated algorithm for further training, evaluation and saving the best model; and S5, inputting the obtained basic features and derived features of the pictures into the saved best model to obtain a prediction result.

2. The multi-algorithm fusion image registration method based on integrated algorithm according to claim 1, characterized in that, In S4, specifically comprising: feeding the constructed samples into different integrated algorithms including random forest, LGB and GBDT respectively, confirming the best model parameters under each integrated algorithm according to cross-validation, saving the best algorithm under each integrated algorithm, and finally selecting the best integrated algorithm as the final model for image comparison by evaluating the effect of the test set.

3. The multi-algorithm fusion image registration method based on integrated algorithm according to claim 2, characterized in that, In S5, specifically comprising: inputting the obtained basic features and derived features according to multiple algorithms into the best integrated algorithm model to obtain the final fusion score result, and considering it as the same object if it exceeds the first probability threshold value of the model.

4. A multi-algorithm fusion image matching system based on ensemble algorithm, characterized in that, The method comprises the following steps: an acquisition module for acquiring a picture set; a division module for dividing pictures of the same object in the picture set into a static set and a dynamic set respectively; An association module is configured to assign an association actorId to pictures of the same object, and determine a label label of a sample based on the association actorId; An extraction module is configured to extract picture feature vectors of the static set and the dynamic set respectively based on a preset algorithm, to obtain basic features and derived features; A sample construction module is configured to construct a sample based on the obtained basic features, the derived features, and the association actorId; A training module is configured to input the sample into a preset integrated algorithm for further training, evaluation, and saving of an optimal model; A prediction module is configured to input the obtained basic features and derived features of a picture into the saved optimal model, to obtain a prediction result.

5. An electronic device, comprising: one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1-3.

6. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the method of any one of claims 1-3.

Images