TWI856861B - Document Identification and Verification System - Google Patents

Abstract

A document recognition and verification system includes a processing device, a communication device and a storage device electrically connected to the processing device. The storage device stores a plurality of business rules corresponding to a plurality of document types. The processing device establishes a connection with a data source end device through the communication device to receive a document image including a quick response code corresponding to a document type, and corrects the document image to generate a corrected document image, and performs image recognition on the corrected document image to obtain a document content of the corrected document image and a data content of the quick response code. The processing device selects a business rule corresponding to the corresponding document type according to the data content to verify the document content, and returns a verification result to the data source end device.

Description

Document Identification and Verification System

The present invention relates to a document identification and verification system, and more particularly to a document identification and verification system using a quick response code.

In existing banking or insurance financial services, it is generally necessary to identify and verify the data content of various paper documents. For example, the autopay authorization form submitted by the customer needs to identify the policy number, the insured, whether the receipt is mailed, and the account name, ID number, and debit account of the authorizer. In addition, it is also necessary to verify the seal pattern in the paper document. These operations are a large and complicated processing process for each company. Therefore, how to use an automated system to simplify the process and improve execution efficiency has become an important issue.

Therefore, the purpose of the present invention is to provide a document recognition and verification system that uses quick response codes to simplify the process and improve execution efficiency.

Therefore, one aspect of the present invention is to provide a document identification and verification system, which is applicable to a data source device and includes a communication device, a storage device, and a processing device. The communication device is used to provide a networking function. The storage device stores a plurality of business rules corresponding to a plurality of document types.

The processing device is electrically connected to the communication device and the storage device, and establishes a connection with the data source device through the communication device to receive a document image from the data source device. The document image includes a quick response code (QR code) corresponding to one of the document types. The processing device performs rotation and scaling correction on the document image to generate a corrected document image, and performs image recognition on the corrected document image to obtain a document content of the corrected document image and a data content of the quick response code.

The processing device selects one of the document types that matches the data content, and checks the document content according to the business rule corresponding to the document, and returns a check result to the data source device via the communication device.

In some implementations, the processing device extracts a plurality of feature points from the document image and first performs rotation correction on the document image based on the feature points.

In some embodiments, the storage device further stores a positional relationship between positioning marks of a quick response code. The quick response code includes three positioning marks located at the upper left, upper right, and lower left, respectively. The processing device calculates the distance between any two of the positioning marks based on the positions of the positioning marks and the positional relationship between the positioning marks of the quick response code, so as to perform a proportional scaling correction on the rotated document image, thereby generating the corrected document image that can restore the size ratio of a paper document corresponding to the document image.

In some implementations, the processing device performs image recognition on the corrected image using optical character recognition (OCR) technology, and also returns the document content to the data source device via the communication device.

In some implementations, the multiple verification items corresponding to each of the business rules include at least an identity card number and a document serial number.

In some implementations, the check items corresponding to each of the business rules further include at least one check content of the document type corresponding to the business rule.

In some implementations, the storage device also stores an object detection model. The processing device detects a seal position in the corrected image by means of the object detection model, and captures a seal image to be authenticated at the seal position. When the processing device determines that the document content meets each of the corresponding verification items to indicate that the verification is correct, a system seal image is obtained by means of the communication device, and the seal image to be authenticated is scaled in proportion to its length and width to be equal to the length of one side of the system seal image, and then overlapped with the system seal image to generate a seal comparison image. The processing device also transmits the seal comparison image back to the data source device via the communication device.

Therefore, another aspect of the present invention is to provide a document identification and verification system, which is applicable to a data source terminal device and includes a communication device and a processing device. The communication device is used to provide a networking function.

The processing device is electrically connected to the communication device, and establishes a connection with the data source device through the communication device to receive a document image from the data source device. The document image includes a quick response code corresponding to a document type. The processing device performs rotation and scaling correction on the document image to generate a corrected document image, and performs image recognition on the corrected document image to obtain a document content of the corrected document image and a data content of the quick response code. The data content includes the document type and at least one business rule corresponding to the document type.

The processing device checks the document content according to the at least one business rule of the data content, and returns a check result to the data source device via the communication device.

In some implementations, the processing device extracts a plurality of feature points from the document image and first performs rotation correction on the document image based on the feature points.

In some embodiments, the storage device further stores a positional relationship between positioning marks of a quick response code. The quick response code includes three positioning marks located at the upper left, upper right, and lower left, respectively. The processing device calculates the distance between any two of the positioning marks based on the positions of the positioning marks and the positional relationship between the positioning marks of the quick response code, so as to perform a proportional scaling correction on the rotated document image, thereby generating the corrected document image that can restore the size ratio of a paper document corresponding to the document image.

The effect of the present invention is that: the processing device first performs rotation and scaling correction on the document image, and then performs image recognition on the corrected document image to obtain the document content and the data content of the quick response code, and selects the document type stored in the storage device that matches the data content, and verifies the document content according to the business rules corresponding to the document type, so as to return the generated verification result to the data source device via the communication device, thereby realizing a document recognition and verification system with simplified process and improved execution efficiency.

Before the present invention is described in detail, it should be noted that similar components are represented by the same reference numerals in the following description.

Referring to FIG. 1 , an embodiment of the document identification and verification system of the present invention is applicable to a data source terminal device 4, and includes a communication device 3, a storage device 2, and a processing device 1. The communication device 3 is used to provide a networking function, and is, for example, a chip or device that supports wired network technology (such as Ethernet) or wireless network technology (such as Wi-Fi), but is not limited thereto. The storage device 2 is, for example, a hard disk, a memory, or other storage device, and stores a plurality of business rules corresponding to a plurality of file types, a position relationship between quick response code location tags, and an object detection model. In addition, in other embodiments, the storage device 2 may also be a storage server, and the method of establishing a connection with the communication device 3 is changed, which will not affect the technical means applied for in this case.

In this embodiment, the document recognition and verification system belongs to an insurance company, for example, and the data source device 4 is a computer device of a person in charge of the insurance company. When the person in charge wants to verify the content of a paper document, the source device takes a photo or scans the paper document to generate a document image, and then transmits the document image to the document recognition and verification system.

The processing device 1 is, for example, one or more central processing units, and is electrically connected to the communication device 3 and the storage device 2. When the processing device 1 establishes a connection with the data source device 4 via the communication device 3, it receives the document image from the source device. The document image includes a quick response code (QR code) corresponding to one of the document types. The quick response code includes three positioning marks located at the upper left, upper right, and lower left, respectively, and all in the shape of a "return". The positional relationship between the quick response code positioning marks stored in the storage device 2 is the relative position and distance of the three positioning marks in the protocol or standard. The processing device 1 performs rotation and scaling correction on the document image to generate a corrected document image.

In more detail, the processing device 1 extracts a plurality of feature points from the document image, and performs rotation correction on the document image based on the feature points. That is, the processing device 1 uses feature matching to identify the features of an object, that is, to identify an object (such as an icon, image, person, etc.) in the document image rather than the entire image, so it is necessary to extract features called "local features". The method is to first select important feature points in the image, and then use the feature points as a basis to describe and compare the surrounding features (local features) to perform rotation. That is, keypoint detection is performed in sequence to obtain keypoints of interest in the image, feature extraction is performed to extract the local features of the area for each keypoint, and then the keypoints are screened and feature matching is performed. The algorithm used for keypoint detection is any one of Fast Feature-Detector (FAST), StarFeatureDetector (STAR), SIFT (nonfree module), SURF (nonfree module), ORB, BRISK, MSER, GoodFeaturesToTrackDetector (GFTT), GoodFeaturesTo- TrackDetector with Harris detector enabled (HARRIS), DenseFeatureDetector (Dense), SimpleBlobDetector (Simple- Blob), etc. The algorithm used for feature point description is any one of Scale Invariant Feature Transform (SIFT), Speed-Up Robust Feature (SURF), Histogram of Oriented Gradient (HOG), Gradient Location Orientation Histogram (GLOH), PCA-SIFT, RootSIFT, Pyramidal HOG (PHOG), Pyramidal Histogram Of visual Words (PHOW), etc. However, the above algorithms are not limited to the above algorithms.

The processing device 1 calculates the distance between any two of the positioning marks based on the positions of the positioning marks in the rotated document image and the positional relationship between the quick response code positioning marks, so as to perform proportional scaling correction on the rotated document image and generate the corrected document image that can restore the size ratio of the paper document corresponding to the document image.

Next, the processing device 1 performs image recognition on the corrected image by optical character recognition (OCR) technology to obtain a document content of the corrected document image, and reads a data content of the quick response code by image recognition technology. For example, the optical character recognition technology adopted is a character recognition software. The data content is the content of the maximum data capacity that can be stored in the corresponding version 40 (such as the document type or information corresponding to the document type). The document content includes an automatic transfer payment authorization form, a policy number, an insured person, whether to mail, the authorizer information (such as account name, ID number, post office information, bank information), etc.

The processing device 1 selects one of the file types stored in the storage device 2 and matching the data content, and checks the file content according to the business rule corresponding to the file to generate a check result. Specifically, the multiple check items corresponding to each business rule at least include an identity card number and a file serial number, and the check items corresponding to each business rule also include at least one check content of the file type corresponding to the business rule. For example, when the document type corresponds to an autopay payment authorization form, the verification contents include whether the policy number conforms to the company format, verification of the ID number, and other corresponding verifications of the insured, whether it is mailed, the information of the authorizer (such as account name, ID number, post office information, bank information), etc. When the processing device 1 determines that the document content conforms to each of the corresponding verification items, the generated verification result indicates that the verification is correct.

In addition, the processing device 1 also detects a fingerprint position in the corrected image by means of the object detection model. That is, the object detection model uses a "frame" to mark the range of the object in the image content such as the photo or video, and classifies it into what kind of object and the guess probability of it being the object, such as the YOLOv7 algorithm, but not limited thereto. Next, the processing device 1 captures a seal image to be authenticated at the seal authentication position. When the processing device 1 determines that the authentication result indicates that the authentication is correct, the communication device 3 obtains a system seal image, for example, from another database, and scales the seal image to be authenticated in a ratio of length and width to make it equal to the length of one side of the system seal image, and then overlaps it with the system seal image to generate a seal comparison image.

The processing device 1 transmits at least one or all of the verification result, the document content, and the seal comparison image back to the data source device 4 via the communication device 3, so that the person in charge can quickly and accurately obtain relevant information on whether the verification is correct.

It is also worth mentioning that: in other embodiments, the data content of the quick response code may include not only the file type but also at least one business rule corresponding to the file type. Therefore, the processing device 1 does not need to obtain the business rule corresponding to the file type from the storage device 2 according to the file type of the data content. Instead, the processing device 1 directly obtains the at least one business rule from the data content. However, it should be noted that: since the data content of the quick response code has a capacity limitation, for example, the maximum data capacity that can be stored in the version 40 quick response code is 984 Chinese characters or 4296 letters, so the number and number of characters of the at least one business rule included in the data content are also subject to capacity limitations.

In summary, the processing device 1 first performs rotation and scaling correction on the document image, and then performs image recognition on the corrected document image to obtain the document content and the data content of the quick response code, and selects the file type stored in the storage device 2 that matches the data content, and checks the document content according to the business rules corresponding to the file type to generate the verification result. Furthermore, when it is determined that the verification is correct, the verification process can be executed to generate the seal comparison image, so that the data source end device 4 can quickly and correctly obtain the verification result, the document content, and the seal comparison image, thereby realizing a document recognition and verification system with simplified process and improved execution efficiency, so that the purpose of the present invention can be achieved.

However, the above is only an embodiment of the present invention and should not be used to limit the scope of implementation of the present invention. All simple equivalent changes and modifications made according to the scope of the patent application of the present invention and the content of the patent specification are still within the scope of the present patent.

1: Processing device 2: Storage device 3: Communication device 4: Data source device

Other features and functions of the present invention will be clearly presented in the implementation method of the reference drawings, in which: Figure 1 is a block diagram illustrating an implementation example of the document recognition and verification system of the present invention.

1: Processing device

2: Storage device

3: Communication device

4: Data source device

Claims (6)

A document recognition and verification system is applicable to a data source terminal device and includes: a communication device for providing a networking function; a storage device for storing a plurality of business rules corresponding to a plurality of document types, and a position relationship between a quick response code location marker; and a processing device electrically connected to the communication device and the storage device, and establishing a connection with the data source terminal device through the communication device to receive a document image from the data source terminal device, the document image including a quick response code (QR code) corresponding to one of the document types. The quick response code includes three positioning marks located at the upper left, upper right, and lower left, respectively. The processing device performs rotation and scaling correction on the document image to generate a corrected document image, and performs image recognition on the corrected document image to obtain a document content of the corrected document image and a data content of the quick response code, wherein the processing device extracts a plurality of feature points from the document image, and performs rotation correction on the document image according to the feature points, and the processing device performs rotation correction on the document image according to the positions of the positioning marks. , and the positional relationship between the quick response code positioning marks, calculate the distance between any two of the positioning marks, so as to perform proportional scaling correction on the rotated document image, and generate the corrected document image that can restore the size ratio of a paper document corresponding to the document image. The processing device selects one of the document types that matches the data content, and checks the document content according to the business rules corresponding to the document, and transmits a check result back to the data source device via the communication device. The document recognition and verification system as described in claim 1, wherein the processing device performs image recognition on the corrected image by optical character recognition (OCR) technology, and also transmits the document content back to the data source device via the communication device. A document identification and verification system as described in claim 2, wherein the multiple verification items corresponding to each business rule include at least an identity card number and a document serial number. A document identification and verification system as described in claim 3, wherein the verification items corresponding to each business rule further include at least one verification content of the document type corresponding to the business rule. The document recognition and verification system as described in claim 4, wherein the storage device further stores an object detection model, the processing device detects a seal position in the corrected image by means of the object detection model, and captures a seal image to be verified at the seal position, and when the processing device determines that the document content meets the corresponding verification items, the processing device indicates that the document content meets the corresponding verification items. When the verification is correct, a system seal image is obtained through the communication device, and the seal image to be authenticated is scaled in proportion to the length and width to be equal to the length of one side of the system seal image, and then overlapped with the system seal image to generate a seal comparison image. The processing device also returns the seal comparison image to the data source device via the communication device. A document recognition and verification system is applicable to a data source terminal device and includes: a storage device for storing a position relationship between positioning marks of a quick response code; a communication device for providing a networking function; and a processing device for electrically connecting the storage device and the communication device, and establishing a connection with the data source terminal device through the communication device to receive a document image from the data source terminal device, the document image including a quick response code corresponding to a document type, the quick response code including three positioning marks located at the upper left, upper right, and lower left, respectively. The processing device performs rotation and scaling correction on the document image to generate a corrected document image, and performs image recognition on the corrected document image to obtain a document content of the corrected document image and the quick response code. A data content of the response code, the data content includes the document type and at least one business rule corresponding to the document type, wherein the processing device extracts multiple feature points from the document image, and first performs rotation correction on the document image according to the feature points, and the processing device calculates the distance between any two of the positioning marks according to the positions of the positioning marks and the position relationship between the positioning marks of the quick response code, so as to perform proportional scaling correction on the rotated document image, and generate the corrected document image that can restore the size ratio of a paper document corresponding to the document image; the processing device checks the document content according to the at least one business rule of the data content, and returns a check result to the data source end device via the communication device.

Images