Bank flow image recognition method and device, readable storage medium and electronic equipment
Through the bank statement image recognition method, image feature fusion technology is used to solve the problem of low efficiency of manual review of bank statements, and fast and accurate identification and audit are achieved, reducing labor costs.
Patent Information
- Application Number
- CN202510439968.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-09
- Publication Date
- 2025-07-18
AI Technical Summary
The existing bank statement analysis and review mainly rely on manual methods, which are inefficient and consume a lot of labor costs, making it difficult to complete authenticity inspection efficiently and at zero risk.
The bank statement image recognition method is adopted. By acquiring images and extracting data information, performing integrity and rationality tests, the image is divided into multiple blocks, and image feature information from different dimensions is extracted for feature fusion and detection to judge the authenticity of the image block.
It realizes fast and accurate bank statement identification, improves inspection efficiency, and reduces labor costs.
Smart Images
Figure CN120340057A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and particularly to a method, device, readable storage medium, and electronic device for identifying bank statement images. Background Art
[0002] In the risk control field of finance, bank statement analysis is a very important risk control strategy. The existing analysis and review of bank statements are basically carried out manually by observing whether the content is abnormal.
[0003] Bank statements mainly exist in unstructured forms such as paper, pictures, and PDFs. With traditional review methods, it is very difficult to complete the authenticity verification of bank statements with high efficiency and zero risk, and it requires a large amount of human cost. Summary of the Invention
[0004] In view of the above situation, it is necessary to provide a method, device, readable storage medium, and electronic device for identifying bank statement images to solve the problem of low efficiency in existing bank statement verification.
[0005] On the one hand, the present invention discloses a method for identifying bank statement images, including: Obtaining a bank statement image and extracting data information from the bank statement image; Performing integrity and rationality checks on the data information; When the integrity and rationality checks of the data information are qualified, performing an authenticity check on the bank statement image, and the steps of the authenticity check include: Dividing the bank statement image into multiple image blocks, and extracting different-dimensional image feature information for each of the image blocks; Performing feature fusion on the image feature information of each dimension of the image block, and detecting the fused image feature information to determine the authenticity of the image block.
[0006] Further, in the above method for identifying bank statement images, the data information includes account balance and incoming and outgoing funds, and the steps of performing integrity and rationality checks on the data information include: Performing continuity checks, missingness checks, rationality checks on the first digit distribution of the account balance, and rationality checks on the incoming and outgoing funds for the account balance.
[0007] Further, in the above method for identifying bank statement images, the step of extracting different-dimensional image feature information for each of the image blocks includes: Extracting spatial domain features in the image block using a spatial domain information extraction network, and extracting correlation features between adjacent pixels in the image block using a correlation feature extraction network.
[0008] Further, in the above bank statement image recognition method, the step of performing feature fusion on the image feature information of each dimension of the image block and detecting the fused image feature information to determine the authenticity of the image block includes: Fusing the spatial domain feature and the correlation feature of the image block; Inputting the fused image feature information into a classification module to predict a probability value of authenticity, and determining the authenticity of the image block according to the prediction result.
[0009] Further, in the above bank statement image recognition method, the step of fusing the spatial domain feature and the correlation feature of the image block includes: After splicing the spatial domain feature and the correlation feature of the image block through channels, performing a convolution operation using a convolution kernel K with an input channel number of c to obtain a fused feature Z concat , Z concat The calculation formula is as follows: , Wherein, represents channel splicing, represents a convolution operation, X and Y respectively represent the spatial domain feature and the correlation feature, and i represents the number of channels.
[0010] On the other hand, the present invention also discloses a bank statement image recognition device, including: An extraction module, configured to obtain a bank statement image and extract data information in the bank statement image; A data verification module, configured to perform integrity and rationality verification on the data information; An image verification module, when the integrity and rationality verification of the data information is qualified, performing authenticity verification on the bank statement image, and the steps of the authenticity verification include: Dividing the bank statement image into a plurality of image blocks, and extracting image feature information of different dimensions for each of the image blocks; Performing feature fusion on the image feature information of each dimension of the image block, and detecting the fused image feature information to determine the authenticity of the image block.
[0011] Further, in the above bank statement image recognition device, wherein the data information includes account balance and incoming and outgoing funds, and the data verification module is used for: Performing continuity verification, missingness verification on the account balance, as well as rationality verification on the first digit distribution of the account balance and rationality verification on the incoming and outgoing funds.
[0012] Further, in the above bank statement image recognition device, the step of extracting image feature information of different dimensions for each of the image blocks includes: extracting the spatial domain features in the image block by using a spatial domain information extraction network, and extracting the correlation features between adjacent pixels in the image block by using a correlation feature extraction network.
[0013] On the other hand, the present invention also discloses an electronic device, including a memory and a processor, where the memory stores a program, and when the program is executed by the processor, the method described in any one of the above is implemented.
[0014] On the other hand, the present invention also discloses a computer-readable storage medium, on which a program is stored, and when the program is executed by a processor, the method described in any one of the above is implemented.
[0015] The present invention performs detections at two levels on the obtained bank statement image, namely the data layer authenticity level and the image authenticity level. Among them, the data authenticity level is relatively easy to detect, so it can be preferentially inspected. It mainly extracts data from the bank statement image and performs data integrity inspection and rationality inspection. The image authenticity level inspection mainly detects the authenticity of the bank statement image, which is a further detection of the bank statement. First, the bank statement image is divided into image blocks, and image feature information is extracted and fused from different dimensions, and the image authenticity is discriminated according to the fused features. This method can quickly and accurately identify bank statements and improve the inspection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 is a flowchart of the bank statement image recognition method in the first embodiment of the present invention; Figure 2 is a structural block diagram of the bank statement image recognition device in the second embodiment of the present invention; Figure 3 is a structural schematic diagram of the electronic device in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0017] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions from beginning to end. The embodiments described below by referring to the drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation to the present invention.
[0018] Embodiments of the present invention will be clear with reference to the following description and drawings. In these descriptions and drawings, some specific embodiments of the embodiments of the present invention are specifically disclosed to represent some ways of implementing the principles of the embodiments of the present invention. However, it should be understood that the scope of the embodiments of the present invention is not limited thereto. On the contrary, the embodiments of the present invention include all variations, modifications, and equivalents falling within the spirit and scope of the appended claims.
[0019] Please refer to Figure 1 , which is the bank statement image recognition method in the first embodiment of the present invention, including steps S11 to S13.
[0020] Step S11: Obtain a bank statement image and extract data information from the bank statement image.
[0021] Specifically, the bank statement image is an image of a bank statement detail list collected by taking a photo. The detail list generally includes information such as the bank name, business verification code, printing date, and data information. The data information at least includes the account balance and incoming and outgoing funds. In specific implementation, OCR technology can be used to extract the text information and data information contained in the bank statement detail list image.
[0022] Step S12: Perform integrity and rationality checks on the data information.
[0023] Performing integrity and rationality checks on the extracted data information may specifically include: Performing continuity checks, missingness checks, reasonableness checks on the first digit distribution of the account balance, and reasonableness checks on the incoming and outgoing funds for the account balance.
[0024] Specifically, for the daily balance of the bank statement, it is required that the daily transaction data of the account is significantly continuous, and for the cross-day balance check, it is required that the account balance is significantly continuous across days. The missingness check requires that the frequency of missing continuous data is within a certain range. In the reasonableness check, the first digit distribution check of the balance requires that the first digit distribution of the account balance conforms to the natural number distribution P(n)=log(1 + 1 / n). Additionally, as an optional implementation method, for the current balance check, it is required that the interest settlement amount, payment time, and marking method in the last month of each quarter are reasonable, etc. The reasonableness check of the incoming and outgoing funds requires that the amount of incoming and outgoing funds in the account in the recent two months is greater than 0. If the above requirements are met, it can be determined that the integrity and rationality checks of the data information are qualified.
[0025] Step S13: When the integrity and rationality checks of the data information are qualified, perform authenticity checks on the bank statement image. The steps of the authenticity check include: Divide the bank statement image into multiple image blocks, and extract different-dimensional image feature information for each of the image blocks; Fuse the image feature information of each dimension of the image block, and detect the fused image feature information to determine the authenticity of the image block.
[0026] In this embodiment, the collected bank statement images are divided into multiple image blocks, and each image block is detected separately to reduce the computational complexity of each image processing. Further, to improve the detection efficiency and accuracy, overlapping sliding windows can be used to divide the bank statement image into multiple image blocks, so that when analyzing each image block, the possibly abnormal image regions can be detected multiple times to improve the inspection accuracy and avoid missed detections. The size of the sliding window can be set according to actual needs. For example, the size can be set to 128 × 128, and this size can achieve the best analysis efficiency and accuracy.
[0027] Extract image feature information of different dimensions for each image block to capture effective image features. For example, it can capture information such as deformed images or text contours, abnormal contrast or brightness in the image. Further, in one implementation manner of the present invention, the step of extracting image feature information of different dimensions for each of the image blocks includes: Extract the spatial domain features in the image block by using a spatial domain information extraction network, and extract the correlation features between adjacent pixels in the image block by using a correlation feature extraction network.
[0028] That is, extract the spatial domain features and correlation features in the image block. Among them, the spatial domain features can be extracted by designing an encoder based on a pre-trained Convolutional Neural Network (CNN) as the basic structure. The correlation feature is the correlation feature between pixels. Specifically, in implementation, the semantic information in the image block can be filtered out by a residual filter, and the correlation between pixels is retained, so that the abnormality between pixels can be captured. For example, in this embodiment, a pre-trained Correlation Feature Extraction Network (CFEN) can be used to extract the correlation features. The CFEN network model is a deep learning model based on the siamese network architecture, and can be end-to-end trained from scratch offline on multiple large-scale data sets. It mainly includes a residual filter, convolutional layers and pooling layers. Through this residual filter, a feature map containing the correlation of pixel neighborhoods with six channels is obtained, and then convolutional and pooling operations are performed respectively to obtain the correlation features between pixels.
[0029] Further, the step of fusing the image feature information of each dimension of the image block, and detecting the fused image feature information to determine the authenticity of the image block includes: Fuse the spatial domain features and correlation features of the image block; Input the fused image feature information into the classification module to predict the authenticity probability value, and determine the authenticity of the image block according to the prediction result.
[0030] The CNN network and the CFEN network can be used as two branch networks, which are respectively used to extract spatial domain features and correlation features, and the features output by the two branch networks can be fused by channel splicing. Specifically, when implemented, after the features output by the two branch networks are spliced by channels, a convolution kernel K with an input channel number of 2c is used, and the convolution result is Z concat , Z concat The calculation formula is as follows: , where, represents channel splicing, represents the convolution operation, X and Y respectively represent the spatial domain features and correlation features, and i represents the number of channels.
[0031] Z concat The first c channels in Z are obtained by convolving several channels of the spatial domain features with c different convolution kernels, and the last c channels are obtained by convolving several channels of the correlation features with another c different convolution kernels. That is, information extraction and integration are respectively performed on the spatial domain features and correlation features.
[0032] In this embodiment, the classification module includes a global average pooling layer, a fully connected layer, and a softmax layer. The spliced features are input into the average pooling layer for average value calculation, and then mapped to two different values by the fully connected layer. The softmax layer maps the output of the fully connected layer to two probability values, and the probability of the authenticity of the image block can be obtained. The authenticity of the image block is determined according to the authenticity probability value.
[0033] This embodiment performs detections on the obtained bank statement images at two levels, namely the data layer authenticity level and the image authenticity level. Among them, the data authenticity level is relatively easy to detect, so it can be preferentially inspected. It mainly extracts data from the bank statement images and performs data integrity inspection and rationality inspection. The image authenticity level inspection mainly detects the authenticity of the bank statement images, which is a further detection of the bank statement. First, the bank statement image is divided into image blocks, and image feature information is extracted and fused from different dimensions, and the image authenticity is judged according to the fused features. This method can quickly and accurately identify bank statements and improve the inspection efficiency.
[0034] Please refer to Figure 2 , which is the bank statement image recognition device in the second embodiment of the present invention, including: An extraction module 21, configured to obtain a bank statement image and extract data information from the bank statement image; A data verification module 22, configured to verify the integrity and rationality of the data information; An image verification module 23, when the integrity and rationality verification of the data information is qualified, perform authenticity verification on the bank statement image, and the steps of the authenticity verification include: Divide the bank statement image into multiple image blocks, and extract image feature information of different dimensions for each of the image blocks; Fuse the image feature information of each dimension of the image block, and detect the fused image feature information to determine the authenticity of the image block.
[0035] Furthermore, for the above bank statement image recognition device, wherein the data information includes account balance and incoming and outgoing funds, and the data verification module is used for: Perform continuity verification, missingness verification on the account balance, as well as rationality verification on the first digit distribution of the account balance and rationality verification on the incoming and outgoing funds.
[0036] Furthermore, for the above bank statement image recognition device, wherein the step of extracting image feature information of different dimensions for each of the image blocks includes: Extract the spatial domain features in the image block by using a spatial domain information extraction network, and extract the correlation features between adjacent pixels in the image block by using a correlation feature extraction network.
[0037] The bank statement image recognition device provided by the embodiments of the present invention has the same implementation principle and the same technical effects as those of the foregoing method embodiments. For the sake of brief description, for the parts not mentioned in the device embodiments, reference may be made to the corresponding contents in the foregoing method embodiments.
[0038] On the other hand, the present invention also proposes an electronic device. Please refer to Figure 3 , as shown in the electronic device in the embodiments of the present invention, including a processor 10, a memory 20, and a computer program 30 stored in the memory and executable on the processor. When the processor 10 executes the computer program 30, the above-mentioned bank statement image recognition method is implemented.
[0039] Wherein, the electronic device may be, but is not limited to, computer devices such as personal computers and mobile phones. The processor 10 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments, and is used to run the program code stored in the memory 20 or process data, etc.
[0040] Among them, the memory 20 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), magnetic memory, magnetic disk, optical disc, etc. In some embodiments, the memory 20 may be an internal storage unit of the electronic device, such as the hard disk of the electronic device. In other embodiments, the memory 20 may also be an external storage device of the electronic device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a FlashCard, etc. equipped on the electronic device. Further, the memory 20 may also include both an internal storage unit and an external storage device of the electronic device. The memory 20 can be used not only to store application software installed on the electronic device and various types of data, etc., but also to temporarily store data that has been output or will be output.
[0041] Optionally, the electronic device may further include a user interface, a network interface, a communication bus, etc. The user interface may include a display, an input unit such as a keyboard. Optionally, the user interface may also include a standard wired interface and a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch liquid crystal display, and an OLED (Organic Light-Emitting Diode) toucher, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display the information processed in the electronic device and to display a visual user interface. The network interface may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface), which is generally used to establish a communication connection between this device and other electronic devices. The communication bus is used to realize the connection and communication between these components.
[0042] It should be noted that Figure 3 The structure shown does not constitute a limitation on the electronic device. In other embodiments, the electronic device may include fewer or more components than shown in the figure, or combine certain components, or have a different component layout.
[0043] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the bank statement image recognition method as described above. Those skilled in the art can understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a defined sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus (such as a computer-based system, a system including a processor, or other systems that can obtain and execute instructions from the instruction execution system, apparatus), or in combination with these instruction execution systems, apparatuses. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in combination with an instruction execution system, apparatus.
[0044] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection portion (electronic device) having one or more wirings, a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.
[0045] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.
[0046] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.
[0047] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that, for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the invention patent shall be subject to the appended claims.
Claims
1. A method for recognizing bank statement images, characterized in that, Including: Obtain a bank statement image and extract data information from the bank statement image; Perform integrity and rationality checks on the data information; When the integrity and rationality checks of the data information are qualified, perform authenticity checks on the bank statement image. The steps of the authenticity check include: Divide the bank statement image into multiple image blocks, and extract image feature information of different dimensions for each of the image blocks; Fuse the image feature information of each dimension of the image block, and detect the fused image feature information to determine the authenticity of the image block.
2. The bank statement image recognition method according to claim 1, wherein, The data information includes account balance and incoming and outgoing funds. The steps of performing integrity and rationality checks on the data information include: Perform continuity checks, missing checks, rationality checks on the first digit distribution of the account balance, and rationality checks on the incoming and outgoing funds for the account balance.
3. The bank statement image recognition method according to claim 1, wherein The steps of extracting image feature information of different dimensions for each of the image blocks include: Extract the spatial domain features in the image block using a spatial domain information extraction network, and extract the correlation features between adjacent pixels in the image block using a correlation feature extraction network.
4. The bank statement image recognition method according to claim 3, characterized in that The steps of fusing the image feature information of each dimension of the image block, detecting the fused image feature information to determine the authenticity of the image block include: Fuse the spatial domain features and correlation features of the image block; Input the fused image feature information into a classification module to predict the authenticity probability value, and determine the authenticity of the image block according to the prediction result.
5. The bank statement image recognition method according to claim 4, wherein, The steps of fusing the spatial domain features and correlation features of the image block include: After the spatial domain features and correlation features of the image block are spliced through channels, a convolution operation is performed using a convolution kernel K with an input channel number of c to obtain the fused feature Z concat , Z concat The calculation formula is as follows: ， Among them, represents channel concatenation, represents a convolution operation, where X and Y represent spatial domain features and correlation features respectively, and i represents the number of channels.
6. A bank statement image recognition device, characterized in that, Including: An extraction module for obtaining a bank statement image and extracting data information from the bank statement image; A data verification module for performing integrity and rationality checks on the data information; An image verification module for performing authenticity checks on the bank statement image when the integrity and rationality checks of the data information are qualified. The steps of the authenticity check include: Divide the bank statement image into multiple image blocks, and extract image feature information of different dimensions for each of the image blocks; Fuse the image feature information of each dimension of the image block, and detect the fused image feature information to determine the authenticity of the image block.
7. The bank statement image recognition device according to claim 6, characterized in that, The data information includes account balance and incoming and outgoing funds. The data verification module is used for: Perform continuity checks, missing checks, rationality checks on the first digit distribution of the account balance, and rationality checks on the incoming and outgoing funds for the account balance.
8. The bank statement image recognition device according to claim 6, wherein, The steps of extracting image feature information of different dimensions for each of the image blocks include: Extract the spatial domain features in the image block using a spatial domain information extraction network, and extract the correlation features between adjacent pixels in the image block using a correlation feature extraction network.
9. An electronic device, characterized in that, Including a memory and a processor. The memory stores a program, and when the program is executed by the processor, the method described in any one of claims 1-5 is implemented.
10. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, the method described in any one of claims 1-5 is implemented.