CN112633148B - Method and system for detecting authenticity of signature fingerprint - Google Patents
Method and system for detecting authenticity of signature fingerprint Download PDFInfo
- Publication number
- CN112633148B CN112633148B CN202011526836.7A CN202011526836A CN112633148B CN 112633148 B CN112633148 B CN 112633148B CN 202011526836 A CN202011526836 A CN 202011526836A CN 112633148 B CN112633148 B CN 112633148B
- Authority
- CN
- China
- Prior art keywords
- fingerprint
- signature
- true
- signature fingerprint
- coding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1382—Detecting the live character of the finger, i.e. distinguishing from a fake or cadaver finger
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/64—Protecting data integrity, e.g. using checksums, certificates or signatures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1347—Preprocessing; Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1365—Matching; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Security & Cryptography (AREA)
- Health & Medical Sciences (AREA)
- Bioethics (AREA)
- General Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Software Systems (AREA)
- Collating Specific Patterns (AREA)
Abstract
A method for detecting the authenticity of a signature fingerprint comprises the following steps: 1) taking out a signature fingerprint with the size of 500 multiplied by 300 from the documents and contracts based on a matting algorithm; 2) based on a feature extraction network stacked by a plurality of lightweight inclusion modules, carrying out texture feature coding on the signature fingerprints; 3) adopting a block label training strategy to strengthen the coding capacity of local texture features of the signature fingerprint, and using a label extension module to complete the training of an offline feature extraction network; 4) and in an online deployment link, policy fusion is carried out on the thermodynamic diagrams output by the network based on the thermodynamic diagram fusion decision scheme, and the confidence that the final signature fingerprint is true is obtained. And to provide a signature fingerprint authenticity detection system. The signature fingerprint activity detection model obtained by the invention effectively reduces the participation of experts and saves a large amount of labor cost.
Description
Technical Field
The invention relates to knowledge in the fields of image processing, machine learning, deep learning, fingerprint identification and the like, and mainly introduces a method and a system for signature fingerprint processing and true and false detection.
Background
With the development of national economy, the signing of the judicial papers becomes an indispensable step in commercial behaviors, the signature fingerprints are used as the embodiment of personal identities on the judicial papers and often become the basis of the fact of responsibility in civil dispute cases, and the notary can judge the legal effectiveness of the contract according to the fact, so that economic disputes and social problems caused by the conditions of default, imposition, signing and the like are avoided to a certain extent. According to the data related to the ministry of public security, some civil cases exist at present, wherein lawbreakers pretend to be true fingers by adopting fingerprint prostheses, and in such cases, lawbreakers evade contracts, obligations on documents and law clauses fulfillment through actions such as counterfeiting contracts, imposition of names, labels and the like, and even worse, extinct lassos can be implemented. Specifically, in the step of pressing the finger print, a signer of a contract or a document copies the surface of the fingertip part of a hand of a person by using materials such as a seal, silica gel, latex and the like to obtain a fingerprint prosthesis capable of replacing a real finger, and the prosthesis is used for replacing the real finger to implement the finger print collection on the contract or the document. The phenomenon seriously reduces the authority and the public credibility of contracts and documents, and causes huge loss to the property of the citizens; thus, there is a need for an effective means to address the issue of such false fingerprint fraud.
Fingerprint liveness detection is a problem that is widely concerned in the field of fingerprint identification. The existing fingerprint anti-counterfeiting technology mostly finishes the authenticity detection in the process of equipment acquisition or carries out the authenticity detection based on digital images acquired by the equipment; the first living body detection is to detect the true and false of the finger based on the dynamic of the sweat of the fingertip, the blood oxygen characteristic and the like, and the second is to extract the characteristic with the living body texture of the fingerprint on the basis of the digital image of the fingerprint to distinguish the true and false of the fingerprint. The two methods can not solve the problem of detecting the authenticity of the fingerprints of the signatures recorded by inkpad and stamp ink on contracts and documents, and the problems in three aspects need to be solved at present. Firstly, a signature fake finger printing material library based on common use in the market is required to be collected and constructed; secondly, the signature fingerprints on the contracts and the documents need a special preprocessing method to intercept and digitize fingerprint areas; finally, the digitized signature fingerprint is different from the fingerprint picture collected by equipment, and a special technical means is needed to distinguish the authenticity of the signature fingerprint. The invention creatively provides a true and false detection method for solving the signature fingerprints on the contracts and the documents, specifically, the texture characteristics of the signature fingerprints are coded by using a deep learning technology, and the detection of the true and false fingerprints is finally completed by combining a heuristic fusion strategy. The method solves the problems of time and labor waste of the expert signature fingerprint living body identification, and achieves the aim of digital intelligent living body signature fingerprint detection.
Disclosure of Invention
In order to overcome the defects of the prior art, the signature fingerprint anti-counterfeiting technology relates to fingerprint prosthesis material library collection, contract document fingerprint area digitalization and signature fingerprint living body detection, the invention provides a method and a system for detecting a contract, a document signature fingerprint to a fingerprint living body by summarizing the commonly used anti-counterfeiting means on fingerprint acquisition equipment, solves the problem of manual identification of commonly used experts by judicial identification, improves the real effectiveness of the signature fingerprints on the contract and the document, and provides a safe and reliable identity basis for civil dispute judgment.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a method for detecting the authenticity of a signature fingerprint comprises the following steps:
1) taking out a signature fingerprint with the size of 500 multiplied by 300 from the documents and contracts based on a matting algorithm;
2) based on a feature extraction network stacked by a plurality of lightweight inclusion modules, carrying out texture feature coding on the signature fingerprints;
3) adopting a block label training strategy to strengthen the coding capacity of local texture features of the signature fingerprint, and using a label extension module to complete the training of an offline feature extraction network;
4) and in an online deployment link, policy fusion is carried out on the thermodynamic diagrams output by the network based on the thermodynamic diagram fusion decision scheme, and the confidence that the final signature fingerprint is true is obtained.
Further, the processing procedure of the step 1) is as follows:
1.1) scanning contracts and documents by using a scanner to obtain RGB original color pictures containing fingerprints;
1.2) correcting the resolution of the original color picture into 500DPI by using a resolution adjustment algorithm;
1.3) framing an effective area of the fingerprint according to the difference of the inkpad, the stamp-pad ink and the color information of the paper bearing the character information;
1.4) further fine-tuning the effective area manually by using a rectangular frame, and deducting a BMP color picture with the picture size of 500 width by 300 height from the original picture, wherein the standard picture size is 440 KB.
Still further, the processing procedure of the step 3) is as follows:
3.1) construct two thermodynamic diagrams of 11x11, which can also be called probability diagrams, to characterize the detection results of the original input. Wherein each point in the output thermodynamic diagram may encode 13x13 of the original input information;
3.2) carrying out extension coding on the label according to the label information of the input original picture to generate an 11x11 label graph.
Further, the processing procedure of the step 4) is as follows:
4.1) extracting a probability graph of the signature fingerprint coding network output as a true fingerprint, and using the probability graph with the size of 11x11 as the final judgment of the true fingerprint and the false fingerprint;
4.2) designing a heuristic fusion strategy of a probability map to decide the authenticity of the fingerprint, comprising the following steps:
4.2.1) number of tokens N statistically true based on a threshold th ═ 0.5 f And the number of false marks is N g ；
4.2.2) calculating | N f -N g The size of |, Δ N;
4.2.3) when Δ N<At 20, taking the average value of the first 110 probability values as the final true fingerprint confidence coefficient; otherwise, calculating N max ＝max(N f ，Ng)；
4.2.4) when N max >At 100 hours, counting corresponding N max Taking the mean value of the individual probabilities as the final true fingerprint confidence coefficient; otherwise, taking the average of the probabilities of the residual quantity as the final true fingerprint confidence.
A system for realizing a signature fingerprint true and false detection method comprises the following steps:
the pre-processing module of the signature fingerprint, deduct and take out the signature fingerprint of size 500x300 from the file, contract on the basis of the algorithm of matting;
the signature fingerprint texture coding module is used for carrying out texture feature coding on the signature fingerprint based on a feature extraction network stacked by a plurality of lightweight inclusion modules;
the block label extension coding module is used for enhancing the coding capacity of local texture features of the signature fingerprints by adopting a block label training strategy and finishing the training of an offline feature extraction network by using the label extension module;
and the thermodynamic diagram strategy fusion module is used for deploying links on line, performing strategy fusion on the thermodynamic diagrams output by the network based on a thermodynamic diagram fusion decision scheme, and obtaining the confidence coefficient that the final signature fingerprint is true.
The beneficial effects of the invention are as follows: a standardized preprocessing method is provided for the signature fingerprints on contracts and documents, the method realizes the automatic detection of effective fingerprint areas by using an area color detection algorithm, and the labor cost of manually framing fingerprint areas is avoided to a certain extent. Secondly, the block label training strategy provided by the invention is beneficial to local coding of the input network signature fingerprint, and the characterization capability of the network on the local details of the signature fingerprint is enhanced. Finally, the thermodynamic diagram fusion strategy provided by the invention has the effect of fully mining the multiple outputs of the network, and the confidence score obtained after fusion is safer and more reliable. The whole scheme of the invention effectively reduces the participation of experts and saves a large amount of labor cost.
Drawings
FIG. 1 is a flow chart of an implementation of the algorithm of the present invention;
FIG. 2 is a flow chart of pre-processing of a contract, paperwork signature fingerprint;
FIG. 3 is a multi-module stacked network based on the inclusion design;
FIG. 4 is a true-false thermodynamic diagram of a network encoded output;
FIG. 5 is an effect diagram of tag extension encoding;
fig. 6 is a converged decision flow diagram of a network output thermodynamic diagram.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to fig. 1 to 6, a method for detecting the authenticity of a signature fingerprint, taking an offline training link and an online real-time deployment link of a contract, when a signature fingerprint exists on a document as an example, the method comprises the following steps:
1) as shown in fig. 2, according to the characteristics of the signature fingerprints on the contracts and the documents, the invention designs a standardized flow of the signature fingerprints, and a signature fingerprint picture with fixed resolution, fixed size and fixed size can be intercepted from the contracts and the documents through the flow, so that a preprocessing module for forming the signature fingerprints is constructed; the implementation steps comprise:
1.1) scanning contracts and documents by using a scanner to obtain RGB original color pictures containing fingerprints;
1.2) correcting the resolution of the original color picture into 500DPI by using a resolution adjustment algorithm;
1.3) framing an effective area of the fingerprint according to the difference of the inkpad, the stamp-pad ink and the color information of the paper bearing the character information;
1.4) further fine-tuning the effective area manually by using a rectangular frame, and deducting a BMP color picture with the picture size of 500x300 (width x height) from the original picture, wherein the standard picture size is 440 KB.
2) As shown in fig. 3, a framework for fully encoding the texture features of the signature fingerprint picture is constructed based on a multi-stage inclusion module stack, and the framework is composed of three inclusion modules and two meanpo modules, which are referred to as signature fingerprint texture encoding modules in the present invention; the implementation steps are as follows:
2.1) capturing the characteristics of different receptive fields of the original input by using a four-branch extended paradigm according to an inclusion module proposed by GooglLeNet; 2/3 of the number of the original inclusion module convolution layer channels is used as a new inclusion module;
2.2) using MeanPooling to connect three inclusion modules, the construction of the full convolution network body framework is completed.
3) The false fingerprints cause the difference between local information and the true fingerprints in the imitation process; therefore, the invention provides a block label supervision training technology, which can enhance the coding capability of the whole network on the local texture of the input information and obtain stronger characteristics beneficial to activity detection. Specifically, the method comprises the steps of constructing two strategies of thermodynamic diagrams and label extension encoding, such as the processes shown in fig. 4 and 5, wherein the process is a block label extension encoding module of the invention; the implementation steps are as follows:
3.1) based on network characteristics, two thermodynamic diagrams of 11x11 were constructed, one for false fingerprints and the other for true fingerprints. The thermodynamic diagram may also be referred to as a probability map, which characterizes the detection results of the original input. According to the definition of receptive field:
l k ＝l k-1 +(f k -1)*s k
wherein l k The receptive field of each point of the k-th layer,/ k-1 The receptive field of each point of the k-1 st layer, f k Is the convolution kernel size of the k-th layer, s k Step size for the k-th layer. According to the formula, the original input information of 13x13, which can be coded by each point in the output thermodynamic diagram, can be calculated;
3.2) carrying out extension coding on the label according to the label information of the input original picture to generate an 11x11 label graph, and finishing an off-line training link according to the label graph.
4) In the implementation and deployment link, a heuristic fusion strategy is used to decide and synthesize a multi-output result based on an output probability map according to the attached figure 6, and a final result is obtained. The whole process of the stage is called as a thermodynamic diagram strategy fusion module; the implementation steps are as follows:
4.1) extracting a probability graph of the signature fingerprint coding network output as a true fingerprint, and using the probability graph with the size of 11x11 as the final judgment of the true fingerprint and the false fingerprint;
4.2) designing a heuristic fusion strategy of a probability map to decide the authenticity of the fingerprint, and finishing the online deployment of the fingerprint authenticity detection in the process. The implementation steps are as follows:
4.2.1) number of tokens N statistically true based on a threshold th ═ 0.5 f And the number of false marks is N g ；
4.2.2) calculating | N f -N g The size of |, Δ N;
4.2.3) when Δ N<At 20, taking the average value of the first 110 probability values as the final true fingerprint confidence coefficient; otherwise, calculating N max ＝max(N f ，Ng)；
4.2.4) when N max >At 100 hours, counting corresponding N max Taking the mean value of the individual probabilities as the finalTrue fingerprint confidence of; otherwise, taking the average of the probabilities of the residual quantity as the final true fingerprint confidence.
BMP pictures with fixed resolution, fixed size and fixed size can be successfully captured from the contracts and the documents according to the flow; further, the original information of the input network signature fingerprint can be coded to obtain the active characteristics capable of representing the authenticity of the fingerprint; and finally, using strategy fusion to obtain the confidence coefficient of the input network fingerprint picture as the basis for deciding the authenticity of the fingerprint. The process is different from a method for recording fingerprint information by traditional fingerprint acquisition equipment, and forms a method for detecting the authenticity of a signature fingerprint.
The embodiments described in this specification are merely illustrative of implementations of the inventive concepts, which are intended for purposes of illustration only. The scope of the present invention should not be construed as being limited to the particular forms set forth in the embodiments, but is to be accorded the widest scope consistent with the principles and equivalents thereof as contemplated by those skilled in the art.
Claims (3)
1. A method for detecting the authenticity of a signature fingerprint is characterized by comprising the following steps:
1) taking out a signature fingerprint with the size of 500 multiplied by 300 from the documents and contracts based on a matting algorithm;
2) based on a feature extraction network stacked by a plurality of light-weight inclusion modules, performing texture feature coding on the signature fingerprint;
3) adopting a block label training strategy to strengthen the coding capacity of local texture features of the signature fingerprint, and using a label extension module to complete the training of an offline feature extraction network;
4) in an online deployment link, strategy fusion is carried out on the thermodynamic diagrams output by the network based on a thermodynamic diagram fusion decision scheme, and the confidence that the final signature fingerprint is true is obtained;
the treatment process of the step 3) is as follows:
3.1) constructing two thermodynamic diagrams of 11x11, which can also be called probability diagrams, and are used for characterizing detection results of original input, wherein each point in the output thermodynamic diagram can encode the original input information of 13x 13;
3.2) carrying out extension coding on the label according to the label information of the input original picture to generate an 11x11 label graph;
the processing procedure of the step 4) is as follows:
4.1) extracting a probability graph of the signature fingerprint coding network output as a true fingerprint, and using the probability graph with the size of 11x11 as the final judgment of the true fingerprint and the false fingerprint;
4.2) designing a heuristic fusion strategy of a probability map to decide the authenticity of the fingerprint, comprising the following steps:
4.2.1) number of tokens N statistically true based on a threshold th ═ 0.5 f And the number of false marks is N g ；
4.2.2) calculating | N f -N g The size of |, Δ N;
4.2.3) when Δ N<At 20, taking the average value of the first 110 probability values as the final true fingerprint confidence coefficient; otherwise, calculating N max ＝max(N f ，Ng)；
4.2.4) when N max >At 100 hours, counting corresponding N max Taking the mean value of the individual probabilities as the final true fingerprint confidence coefficient; otherwise, taking the average of the probabilities of the residual quantity as the final true fingerprint confidence.
2. The method for detecting the authenticity of the signature fingerprint according to claim 1, wherein the processing procedure of the step 1) is as follows:
1.1) scanning contracts and documents by using a scanner to obtain RGB original color pictures containing fingerprints;
1.2) correcting the resolution of the original color picture into 500DPI by using a resolution adjustment algorithm;
1.3) framing an effective area of the fingerprint according to the difference of the inkpad, the stamp-pad ink and the color information of the paper bearing the character information;
1.4) further fine-tuning the effective area manually by using a rectangular frame, and deducting a BMP color picture with the picture size of 500 width by 300 height from the original picture, wherein the standard picture size is 440 KB.
3. A system implemented by the signature fingerprint true and false detection method of claim 1, wherein the system comprises:
the pre-processing module of the signature fingerprint, deduct and take out the signature fingerprint of size 500x300 from the file, contract on the basis of the algorithm of matting;
the signature fingerprint texture coding module is used for carrying out texture feature coding on the signature fingerprint based on a feature extraction network stacked by a plurality of lightweight inclusion modules;
the block label extension coding module is used for enhancing the coding capacity of local texture features of the signature fingerprints by adopting a block label training strategy and finishing the training of an offline feature extraction network by using the label extension module;
and the thermodynamic diagram strategy fusion module is used for deploying links on line, performing strategy fusion on the thermodynamic diagrams output by the network based on a thermodynamic diagram fusion decision scheme, and obtaining the confidence coefficient that the final signature fingerprint is true.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011526836.7A CN112633148B (en) | 2020-12-22 | 2020-12-22 | Method and system for detecting authenticity of signature fingerprint |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011526836.7A CN112633148B (en) | 2020-12-22 | 2020-12-22 | Method and system for detecting authenticity of signature fingerprint |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112633148A CN112633148A (en) | 2021-04-09 |
| CN112633148B true CN112633148B (en) | 2022-08-09 |
Family
ID=75320635
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011526836.7A Active CN112633148B (en) | 2020-12-22 | 2020-12-22 | Method and system for detecting authenticity of signature fingerprint |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112633148B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103927368A (en) * | 2014-04-22 | 2014-07-16 | 浪潮软件股份有限公司 | Method of lightweight framework for generating thermodynamic diagram according to streaming data concept |
| CN104200239A (en) * | 2014-09-09 | 2014-12-10 | 河海大学常州校区 | Image feature fusion identification based signature authentic identification system and method |
| JP2017064370A (en) * | 2015-09-29 | 2017-04-06 | キヤノン株式会社 | Image processing device, and method and program for controlling image processing device |
| CN110232696A (en) * | 2019-06-20 | 2019-09-13 | 腾讯科技（深圳）有限公司 | A kind of method of image region segmentation, the method and device of model training |
| CN111553916A (en) * | 2020-05-09 | 2020-08-18 | 杭州中科睿鉴科技有限公司 | Image tampering area detection method based on multiple characteristics and convolutional neural network |
| CN111931782A (en) * | 2020-08-12 | 2020-11-13 | 中国科学院上海微系统与信息技术研究所 | Semantic segmentation method, system, medium, and apparatus |
-
2020
- 2020-12-22 CN CN202011526836.7A patent/CN112633148B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103927368A (en) * | 2014-04-22 | 2014-07-16 | 浪潮软件股份有限公司 | Method of lightweight framework for generating thermodynamic diagram according to streaming data concept |
| CN104200239A (en) * | 2014-09-09 | 2014-12-10 | 河海大学常州校区 | Image feature fusion identification based signature authentic identification system and method |
| JP2017064370A (en) * | 2015-09-29 | 2017-04-06 | キヤノン株式会社 | Image processing device, and method and program for controlling image processing device |
| CN110232696A (en) * | 2019-06-20 | 2019-09-13 | 腾讯科技（深圳）有限公司 | A kind of method of image region segmentation, the method and device of model training |
| CN111553916A (en) * | 2020-05-09 | 2020-08-18 | 杭州中科睿鉴科技有限公司 | Image tampering area detection method based on multiple characteristics and convolutional neural network |
| CN111931782A (en) * | 2020-08-12 | 2020-11-13 | 中国科学院上海微系统与信息技术研究所 | Semantic segmentation method, system, medium, and apparatus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112633148A (en) | 2021-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ghiani et al. | Review of the fingerprint liveness detection (LivDet) competition series: 2009 to 2015 | |
| CN101894260B (en) | Method for identifying forgery seal based on feature line randomly generated by matching feature points | |
| Sansone et al. | Signature verification: increasing performance by a multi-stage system | |
| AU2018292176A1 (en) | Detection of manipulated images | |
| Chandra et al. | Offline signature verification based on geometric feature extraction using artificial neural network | |
| CN102542660A (en) | Bill anti-counterfeiting identification method based on bill watermark distribution characteristics | |
| Jain et al. | Indian currency denomination identification using image processing technique | |
| CN112102232B (en) | Method and device for automatically evaluating colony quality of induced pluripotent stem cells | |
| Rao et al. | A novel fingerprints identification system based on the edge detection | |
| Engin et al. | Offline signature verification on real-world documents | |
| Porwik et al. | DTW and voting-based lip print recognition system | |
| Yazdi et al. | A new approach for the fingerprint classification based on gray-level co-occurrence matrix | |
| CN111476727B (en) | Video motion enhancement method for face-changing video detection | |
| Sudharshan et al. | Handwritten signature verification system using deep learning | |
| CN112633148B (en) | Method and system for detecting authenticity of signature fingerprint | |
| CN113807237A (en) | Training of in vivo detection model, in vivo detection method, computer device, and medium | |
| CN111259894B (en) | Certificate information identification method and device and computer equipment | |
| Verma et al. | Offline signature verification and identification using angle feature and pixel density feature and both method together | |
| CN111881880A (en) | Bill text recognition method based on novel network | |
| CN113011284B (en) | Signature fingerprint identification method | |
| CN109359616A (en) | A kind of pseudo- splicing small size algorithm for recognizing fingerprint based on SIFT | |
| Tamrakar et al. | Low resolution fingerprint image verification using CNN filter and LSTM classifier | |
| TWI777188B (en) | Contract signature authentication method and device | |
| Abdala et al. | Offline signature recognition and verification based on artificial neural network | |
| Impedovo et al. | Stability analysis of static signatures for automatic signature verification |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |