CN115862153A - Signature handwriting authentication method integrating point features, local features and global features, storage medium and electronic equipment - Google Patents
Signature handwriting authentication method integrating point features, local features and global features, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN115862153A CN115862153A CN202111124123.2A CN202111124123A CN115862153A CN 115862153 A CN115862153 A CN 115862153A CN 202111124123 A CN202111124123 A CN 202111124123A CN 115862153 A CN115862153 A CN 115862153A
- Authority
- CN
- China
- Prior art keywords
- signature
- strokes
- similarity
- point
- signatures
- 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.)
- Pending
Links
Images
Landscapes
- Collating Specific Patterns (AREA)
Abstract
The invention requests to protect an online signature authentication method of comprehensive point characteristics, local characteristics and global characteristics, relates to the technical field of electronic signature handwriting authentication, and aims to perform time dynamic planning according to point characteristics of a reserved sample signature and a signature to be detected to obtain point characteristic similarity of the reserved sample signature and the signature to be detected, calculate local characteristic similarity of the reserved sample signature and the signature to be detected based on LNPS characteristics, calculate global characteristic similarity of the reserved sample signature and the signature to be detected according to Euclidean distance, calculate comprehensive similarity of the reserved sample signature and the signature to be detected according to the point characteristic similarity, the local characteristic similarity and the global characteristic similarity, and authenticate the signature according to the comprehensive similarity. The method has the advantages of more real and reliable authentication result and improvement of the judgment precision and accuracy in the handwriting recognition and authentication processes.
Description
Technical Field
The invention belongs to the technical field of computer information processing, and particularly relates to a method for comprehensively judging an online authentication signature by using a time dynamic programming method.
Background
Most of the existing online authentication signature methods collect the abscissa, the ordinate and the pressure value of each moment of the user signature, and calculate the point characteristics such as speed, acceleration, angle, angular velocity and the like of each moment according to the information. The invention relates to Chinese patent application Nos. 201310521026, namely an online handwritten signature authentication method based on a dynamic threshold value, and application No. 201610250594, an online signature identity authentication method and system based on mobile phone sensing, which are all based on the point characteristics, and a method for obtaining the similarity between a signature to be tested and a user sample-keeping signature by using a time dynamic programming (DTW) method is adopted, only the strategic change is made on the selection of the threshold value, the distance between the point and the point is calculated by simply depending on the DTW, the distance between corresponding strokes and strokes in the signature is not considered, and the decision judgment is comprehensively carried out. Therefore, the accuracy of the judgment is not high.
The Chinese patent application with publication number CN111461015A 'a user-independent online signature identification method and device based on RNN model' provides a user-independent online signature identification method and device based on RNN model. Respectively splitting the registered signature and the signature to be authenticated according to strokes to obtain a plurality of stroke data, then connecting the stroke data of the registered signature and the signature to be authenticated together, and extracting the connected stroke data by using a stroke RNN model to obtain the stroke characteristics corresponding to the stroke data; then inputting the stroke features into a signature RNN model to extract overall signature features, and calculating values S1 and S2 corresponding to the similarity and the difference of the overall signature features; and calculating according to the values S1 and S2 corresponding to the similarity and the difference of the overall signature characteristics to finally obtain the similarity of the registered signature and the signature to be authenticated, and outputting an authentication result. However, this patent application does not teach how to segment strokes, how to obtain corresponding strokes, and how to obtain stroke data and features. In addition, deep learning is mainly relied on, and the interpretability and the visualization of stroke features are not available.
These point-based methods do not take into account the characteristics of each stroke of the signature and the overall layout and overall structure of the entire signature, which is precisely the most familiar and accepted main feature that distinguishes between authentic and simulated writing, and we consider this to be simulated writing because there is no similarity between certain two strokes, or because these two signatures are surprisingly similar in overall structure. The method of the prior art brings trouble in the precision and accuracy judgment in the handwriting recognition and authentication process.
Disclosure of Invention
Aiming at the problems that the prior art does not fully consider the strokes and the whole structure of the signature, does not consider the relation between the corresponding strokes, does not have interpretability and visualization of the stroke characteristics and the like, the invention provides a DTW-based method for calculating the similarity of the signature based on the point characteristics, and calculating the similarity between the strokes and the similarity of the whole signature structure. The accuracy of judgment is improved.
The DTW-based method calculates the signature similarity based on point features, and calculates the similarity between strokes and the structural similarity of the whole signature. The similarity of the point characteristics, the local characteristics and the global characteristics is integrated, the authenticity of the handwriting is more accurately and effectively identified, and the problems of low accuracy and low efficiency in electronic handwriting identification are solved.
The technical scheme for solving the technical problems is to provide an online authentication signature method of comprehensive point features, local features and global features, which comprises the following steps: and performing time dynamic planning according to the point characteristics of the reserved sample signature and the signature to be detected to obtain the characteristic similarity of the reserved sample signature and the signature to be detected, calculating the local characteristic similarity of the reserved sample signature and the signature to be detected based on LNPS characteristics, calculating the global characteristic similarity of the reserved sample signature and the signature to be detected according to Euclidean distance, calculating the comprehensive similarity of the reserved sample signature and the signature to be detected according to the point characteristic similarity, the local characteristic similarity and the global characteristic similarity, and authenticating the signature according to the comprehensive similarity.
The authenticating the signature according to the comprehensive similarity specifically includes: and carrying out weighted average on the similarity of the local features and the global features to calculate the comprehensive similarity of the reserved sample signature and the signature to be tested, if the similarity is greater than a threshold value, confirming that the signature to be tested is signed by the reserved sample signer, and the authentication is passed, otherwise, the authentication is not passed, wherein the threshold value is the maximum distance or the minimum similarity of the reserved sample signatures of different times.
Collecting the abscissa, ordinate, pressure value, speed, acceleration, angle, angular velocity and the like corresponding to the strokes of the electronic signature at all the moments, taking the value corresponding to each moment as a point feature, and dividing the sample-left signature into a plurality of strokes according to the angular velocity in the point feature. And comparing the angular speed of all the point characteristics of the sample-preserved signature stroke, if the angular speed at a certain moment is greater than a preset angular threshold value, taking the point corresponding to the moment as an end point of the sample-preserved signature stroke to obtain the end points of all the sample-preserved signature strokes, and dividing the sample-preserved signature into a plurality of strokes according to the end points. And performing time dynamic programming (DTW) operation according to the point characteristics of the reserved sample signature and the signature to be detected to obtain the point characteristic similarity of the reserved sample signature and the signature to be detected, namely obtaining the corresponding endpoint of the signature to be detected according to the stroke endpoint of the reserved sample signature and the corresponding relation between the two signature points and the point, thereby obtaining a plurality of strokes corresponding to the reserved sample signature and the signature to be detected one by one based on the point characteristic similarity.
The similarity of the local characteristics of the reserved sample signature and the signature to be detected can be calculated by adopting an LNPS characteristic extraction method. Calculating the track length of strokes according to the angular speed of the strokes of the signature, determining LNPS characteristics of the strokes according to the track length, normalizing a norm of the difference between the LNPS characteristics of the reserved signature and the strokes of the signature to be detected by using the maximum value to obtain the distance of the corresponding strokes, wherein the local characteristic similarity of the two signatures is the weighted average of the distances of all the corresponding strokes of the two signatures, and the weight is the total point number corresponding to the strokes. The method specifically comprises the following steps:
the angular velocity v of the signature stroke in the x-direction and the y-direction depending on the time t x (t),v Y (t) calling formula
Calculating the track length L (X) of the stroke segment X of the signature to be detected, and according to a formula:
calculating LNPS (X) characteristics of the signature stroke segment X, and normalizing a norm of LNPS characteristic difference by a maximum value according to a formula: />
And calculating the distance D (X, X ') between the sample reserving signature stroke X ' and the corresponding to-be-detected signature stroke X, wherein LNPS (X) in the formula is the LNPS characteristic of the to-be-detected signature, LNPS (X ') is the LNPS characteristic of the sample reserving signature, a lower corner mark 1 represents a norm, and max is a maximum value.
And obtaining global feature similarity of the reserved sample signature and the to-be-detected signature according to the Euclidean distance between the to-be-detected signature and the reserved sample signature, replacing the stroke position of the signature with the center point of the signature, calculating the Euclidean distance between any two strokes in the same signature, determining the Euclidean distance between the reserved sample signature and the to-be-detected signature, and determining the global feature similarity according to the Euclidean distance. The method specifically comprises the following steps:
acquiring a central point m (X) and a length L (X) of a track X (t) of a signature stroke X, and calling a formula: l (X, Y) = | | | | m (X) -m (Y) | calculation of eyes 2 And calculating the Euclidean distance between any two strokes X and Y in the signature to be detected. In the same way, the Euclidean distance between two corresponding strokes of the sample retention signature is calculated to be L (X ', Y'), and the formula is called:
and calculating the Euclidean distance between the reserved sample signature and the signature to be detected. The global feature similarity of the two signatures is a weighted average of Euclidean distances between the two signatures according to a formula
And determining similarity weight. Wherein, L (X) and L (Y) are the track lengths of any two segments of strokes X and Y in the signature to be detected, and L (X ') and L (Y') are the track lengths of any two segments of strokes X 'and Y' in the sample-keeping signature.
And finally, calculating the comprehensive similarity of the sample-remaining signature and the signature to be tested by using the point characteristics, the local characteristics and the global characteristics in a weighted average mode. If the similarity is greater than the threshold value, the signature to be detected is signed by the sample-reserving signer; otherwise, it is a fake signature. Wherein the threshold value can be determined as the maximum distance or the minimum similarity of the leave-sample signatures at different times during verification.
The present invention also claims a computer readable storage medium having stored thereon a computer program that can be loaded and executed by a processor to perform the above-mentioned online signature authentication method.
The present invention also claims an electronic device, including: one or more processors; a memory; one or more application programs stored in the memory and configured to be loaded and executed by the one or more processors to perform the above-described online signature authentication method.
The invention comprehensively considers point characteristics, local characteristics and global characteristics in the signature, has more real and reliable authentication results than the traditional on-line signature authentication system, improves the accuracy and the precision of judgment in the processes of handwriting identification and authentication, and can more intuitively display the authentication results of the signature to be detected because the local characteristics and the global characteristics have more interpretability and visualization than the point characteristics and also provide visual interpretability like the traditional manual authentication.
Drawings
FIG. 1 is a flow chart of the present invention for on-line authentication based on the whole feature of a signature stroke.
Detailed Description
The following detailed description of the embodiments of the invention refers to the accompanying drawings and specific examples. The following examples are only used to illustrate the technical solutions of the present invention more clearly, and the protection scope of the present application is not limited thereby. Based on the embodiments of the invention.
FIG. 1 is a flow chart of the present invention for on-line authentication based on the whole signature stroke feature. The method specifically comprises the following steps: collecting characteristics such as abscissa, ordinate and pressure value of a reserved sample signature and a signature to be detected at each moment in the signature process, calculating the speed, acceleration, angle and angular speed of each point in a signature stroke, and taking the values as point characteristics; since two adjacent strokes usually have different inclinations, this results in a distinct angular change of the end point connecting the two strokes, i.e. the angular velocity of the point is significantly greater than the other points, so that the strokes are segmented with an angular velocity having a distinct angular change in the point feature. Dividing the reserved sample signature into a plurality of strokes according to the angular speed, comparing the angular speed of each point in the reserved sample signature strokes, if the angular speed of a certain point is greater than a preset angular threshold (the angular threshold is preset), taking the point as the endpoint of the reserved sample signature strokes to obtain all the endpoints, and dividing the reserved sample signature into a plurality of strokes according to the endpoint.
And performing time dynamic programming (DTW) operation according to the point characteristics of the reserved sample signature and the signature to be detected to obtain the distance between the point characteristics of the two signatures, and obtaining the similarity of the reserved sample signature and the signature to be detected based on the point characteristics and the point-point corresponding relation between the two signatures. The end point of the signature to be tested corresponding to the end point of the reserved sample is obtained according to the end point of the stroke of the reserved sample signature and the corresponding relation between the point and the point between the reserved sample signature and the signature to be tested. Therefore, the similarity of a plurality of strokes corresponding to the reserved sample signature and the signature to be detected one by one based on the point features is obtained. In other words, corresponding stroke end points are obtained according to the corresponding relation between the points obtained by the DTW algorithm, and the strokes corresponding to the sample-keeping signature and the signature to be tested one by one are determined according to the corresponding stroke end points.
The LNPS (Signature profile features normalized by stroke length) feature extraction method in Current Adaptation Networks for Online Signaling Verification can be used. And extracting local features of each stroke of the signature to be detected, and calculating the similarity of the sample-remaining signature and the local features of the signature to be detected. The details are further described as follows:
obtaining the track of a signature stroke segment at each moment, calculating the speed of the signature stroke at each moment, determining the track length of the segment of the stroke, establishing a second-order equation on the segment of the stroke track according to the instantaneous speeds of any two moments in the x direction and the y direction, normalizing the second-order equation by using the track length to obtain the LNPS characteristic of the segment of the stroke, normalizing a norm of the difference between the LNPS characteristic of the corresponding stroke of the reserved signature and the corresponding stroke of the signature to be detected, then normalizing by using the maximum value to obtain the distance between the sample signature and the corresponding stroke of the signature to be detected, and obtaining the local characteristic of the two signatures by weighted averaging the distances between all the corresponding strokes of the two signatures.
Specifically, the trace of the signature stroke segment is obtained according to the abscissa and the ordinate of all the signatures at all the time, the abscissa x (t) and the ordinate y (t) in the signature stroke at the time t are collected, and the trace x (t) = (x (t)) of the signature stroke segment is obtained，y(t)) T T is more than or equal to 0 and less than or equal to T, wherein the signature time of the stroke X is O-T moment. Calculating the instantaneous speed V (t) of the signature stroke at the time t according to the speeds of the abscissa and the ordinate directions at the time t, wherein V (t) = [ V = x (t)，v y (t)] T T is more than or equal to 0 and less than or equal to T; taking the derivative of the square root of the speed in the x-direction and the y-direction, calling the formula:
and calculating the track length L (X) of the stroke X. The velocity integration in the X-direction and y-direction at any two times during the 0-T period establishes a second-order feature on the stroke trajectory X: />
The values in this feature represent the area projection of the signature stroke in xx, xy, yx, yy directions/planes, respectively. Normalizing the second-order characteristic by using the stroke track length L (X) of the segment to obtain the LNPS characteristic of the segment of the stroke:
the distance between the corresponding strokes of the LNPS (X') and LNPS (X) characteristics of the sample signature and the strokes of the signature to be detected, which are obtained according to the method, is a norm of the LNPS characteristic difference, and then the maximum value is used for normalization, namely a formula is called:
and calculating the distance D (X, X') between the sample signature and the stroke of the corresponding segment of the signature to be detected. The local feature similarity of the two signatures is the weighted average of the strokes corresponding to the two signatures, and the weight is the point number of the corresponding stroke.
And (3) extracting global features of the sample-remaining signature and the signature to be detected, specifically, replacing the position of each stroke by a central point, and calculating the Euclidean distance between every two strokes. And determining the similarity of the global features according to the Euclidean distance.
Determining central points m (X) and m (Y) of the tracks X (t) and Y (t) of any two strokes X and Y in the signature to be detected, and calling a formula: l (X, Y) = | | | m (X) -m (Y) | non-woven hair 2 And calculating the Euclidean distance L (X, T) between the X stroke and the Y stroke.
Similarly, the Euclidean distance between two strokes X 'and Y' corresponding to the leave-sample signature is calculated according to the formula, so that the formula is called:
and calculating the Euclidean distance between the sample-remaining signature and the stroke segment corresponding to the signature to be detected. Completing calculation of Euclidean distances between the reserved sample signature and all the stroke segments of the signature to be detected, and obtaining the global characteristics of the two signatures by weighted averaging of the Euclidean distances of all the stroke segments of the two signatures, wherein the final weight is
L (X) and L (y) are the track lengths of any two strokes in the signature to be detected, and L (X ') and L (y') are the track lengths of any two strokes in the sample-keeping signature.
And finally, carrying out weighted average calculation on the point features, the local features and the global features to calculate the comprehensive similarity of the sample-remaining signature and the signature to be detected, if the similarity is greater than a threshold value, judging that the signature to be detected is signed by a sample-remaining signer, and if not, signing by a non-sample-remaining signer to finish authentication. And determining the maximum distance or the minimum similarity of the sample retention signatures of different times in the verification as a threshold value.
The invention provides an online signature authentication method of comprehensive point characteristics, local characteristics and global characteristics, which is characterized by comprising the steps of performing time dynamic planning according to point characteristics of a reserved sample signature and a signature to be tested to obtain point characteristic similarity of the reserved sample signature and the signature to be tested, calculating the local characteristic similarity of the reserved sample signature and the signature to be tested based on LNPS characteristics, calculating the global characteristic similarity of the reserved sample signature and the signature to be tested according to Euclidean distance, calculating the comprehensive similarity of the reserved sample signature and the signature to be tested according to the point characteristic similarity, the local characteristic similarity and the global characteristic similarity, and authenticating the signature according to the comprehensive similarity. The method has the advantages of more real and reliable authentication result and improvement of the judgment precision and accuracy in the handwriting recognition and authentication processes. The authentication method is implemented by a computer program that can be loaded and executed by a processor to perform the online signature authentication method described above. Also, the electronic device may be an electronic device comprising a plurality of processors, a memory, and an application program, and may be configured to be loaded and executed by the one or more processors so as to perform the above-described online signature authentication method.
Claims (9)
1. An electronic signature handwriting authentication method integrating point features, local features and global features is characterized by comprising the following steps of: and performing time dynamic planning according to the point characteristics of the reserved sample signature and the signature to be detected to obtain the point characteristic similarity of the reserved sample signature and the signature to be detected, calculating the local characteristic similarity of the reserved sample signature and the signature to be detected based on LNPS characteristics, calculating the global characteristic similarity of the reserved sample signature and the signature to be detected according to Euclidean distance, calculating the comprehensive similarity of the reserved sample signature and the signature to be detected according to the point characteristic similarity, the local characteristic similarity and the global characteristic similarity, and authenticating the signature according to the comprehensive similarity.
2. The method of claim 1, wherein authenticating the signature based on the integrated similarity specifically comprises: and carrying out weighted average on the similarity of the local features and the global features to calculate the comprehensive similarity of the reserved sample signature and the signature to be tested, if the similarity is greater than a threshold value, confirming that the signature to be tested is signed by the reserved sample signer, and the authentication is passed, otherwise, the authentication is not passed, wherein the threshold value is the maximum distance or the minimum similarity of the reserved sample signatures of different times.
3. The method as claimed in claim 1 or 2, wherein the point features are abscissa, ordinate, pressure value, velocity, acceleration, angle, angular velocity values corresponding to each moment in the strokes of the electronic signature, angular velocity values in all the point features of the reserved sample signature stroke are compared, if the angular velocity at a certain moment is greater than a predetermined angular threshold, the point corresponding to the moment is used as an end point of the reserved sample signature stroke, the reserved sample signature is divided into a plurality of strokes according to the end point, the end point corresponding to the signature to be tested is obtained according to the reserved sample signature stroke and the corresponding relationship between the reserved sample signature and the point of the signature to be tested, and the similarity between the reserved sample signature and the strokes corresponding to the signature to be tested is obtained based on the point features.
4. The method of claim 1 or 2, wherein calculating the local feature similarity of the proof signature and the signature to be tested further comprises: calculating the track length of the strokes according to the angular speed of the strokes of the signature, determining LNPS characteristics of the strokes according to the track length, normalizing a norm of a difference between the LNPS characteristics of the strokes of the reserved signature and the to-be-detected signature by using a maximum value to obtain the distance between the strokes corresponding to the two signatures, carrying out weighted average on the distances between all the corresponding strokes of the two signatures to obtain the local characteristic similarity of the two signatures, and taking the weight as the total number of points corresponding to the strokes.
5. Method according to claim 4, characterized in that the speed v of the test signature stroke in x-direction and y-direction according to the time t x (t),v Y (t) calling formula
Calculating the track length L (X) of the to-be-detected signature stroke segment X, and forming a second-order characteristic I according to the area projection of the signature stroke on xx, xy, yx and yy 2 (x) Calling a formula:
calculating LNPS (X) characteristics of the to-be-detected signature stroke segment X, calculating LNPS (X ') characteristics of the reserved sample signature stroke segment X' by the same method, and according to a formula: />
And normalizing the norm of the LNPS characteristic difference of the two signatures by using the maximum value to obtain the distance D (X, X ') between the strokes corresponding to the two signatures, wherein the LNPS (X) in the formula is the LNPS characteristic of the signature to be detected, the LNPS (X') is the LNPS characteristic of the reserved sample signature, the lower corner mark 1 represents a norm, and max is the maximum value.
6. The method of any one of claims 1-4, wherein obtaining global feature similarity of two signatures further comprises: the method comprises the steps of obtaining the central point of strokes of the signature, calculating Euclidean distance between any two strokes in the same signature according to the central point, normalizing a norm of Euclidean distance difference of the strokes corresponding to the two signatures by using the maximum value to obtain the Euclidean distance of the strokes corresponding to the two signatures, and carrying out weighted average on the Euclidean distances of all the strokes corresponding to the two signatures to obtain the global feature similarity of the two signatures.
7. The method as claimed in claim 6, wherein the central points m (X), m (Y) of the X and Y traces of any two stroke segments in the signature to be tested are obtained, and the formula is called: l (X, Y) = | | | | m (X) -m (Y) | calculation of eyes 2 Calculating Euclidean distance between the stroke segments X and Y of the signature to be detected, calculating and obtaining Euclidean distance L (X ', Y') between two corresponding strokes of the reserved sample signature by the same method, and calling a formula:
normalizing a norm of the Euclidean distance difference of the strokes corresponding to the two signatures by using a maximum value to obtain the Euclidean distance of the strokes corresponding to the two signatures, wherein the weighted average of all the Euclidean distances of the corresponding strokes between the two signatures is the global feature similarity of the two signatures, and the weight of the similarity is ^ or ^>
L (X) and L (Y) are the track lengths of any two strokes X and Y in the signature to be tested, and L (X ') and L (Y') are the track lengths of any two strokes X 'and Y' in the sample-remaining signature.
8. A computer-readable storage medium, having stored thereon a computer program which can be loaded and run by a processor to perform the online signature authentication method of any one of claims 1 to 7.
9. An electronic device, comprising: one or more processors; a memory; one or more application programs stored in the memory and configured to be loaded and executed by the one or more processors to perform the online signature authentication method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111124123.2A CN115862153A (en) | 2021-09-24 | 2021-09-24 | Signature handwriting authentication method integrating point features, local features and global features, storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111124123.2A CN115862153A (en) | 2021-09-24 | 2021-09-24 | Signature handwriting authentication method integrating point features, local features and global features, storage medium and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115862153A true CN115862153A (en) | 2023-03-28 |
Family
ID=85652612
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111124123.2A Pending CN115862153A (en) | 2021-09-24 | 2021-09-24 | Signature handwriting authentication method integrating point features, local features and global features, storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115862153A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117592125A (en) * | 2024-01-19 | 2024-02-23 | 湖南省不动产登记中心 | High-reliability electronic signature method of paperless transaction system for second-hand house transaction |
-
2021
- 2021-09-24 CN CN202111124123.2A patent/CN115862153A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117592125A (en) * | 2024-01-19 | 2024-02-23 | 湖南省不动产登记中心 | High-reliability electronic signature method of paperless transaction system for second-hand house transaction |
| CN117592125B (en) * | 2024-01-19 | 2024-04-09 | 湖南省不动产登记中心 | Reliability electronic signature method of paperless transaction system for second-hand house transaction |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20140147023A1 (en) | Face Recognition Method, Apparatus, and Computer-Readable Recording Medium for Executing the Method | |
| JP3335538B2 (en) | Method and apparatus for collating a handwritten character string | |
| CN105574518A (en) | Method and device for human face living detection | |
| CN103927532B (en) | Person's handwriting method for registering based on stroke feature | |
| CN105912910A (en) | Cellphone sensing based online signature identity authentication method and system | |
| EP2360619A1 (en) | Fast fingerprint searching method and fast fingerprint searching system | |
| US11132531B2 (en) | Method for determining pose and for identifying a three-dimensional view of a face | |
| US20190121951A1 (en) | A method for online signature verification using wrist-worn devices | |
| Ramachandra et al. | Robust offline signature verification based on global features | |
| CN111178203B (en) | Signature verification method and device, computer equipment and storage medium | |
| CN115862153A (en) | Signature handwriting authentication method integrating point features, local features and global features, storage medium and electronic equipment | |
| CN100452082C (en) | Signature identifying method | |
| Madabusi et al. | On-line and off-line signature verification using relative slope algorithm | |
| CN106980845B (en) | Face key point positioning method based on structured modeling | |
| Tolosana et al. | Increasing the robustness of biometric templates for dynamic signature biometric systems | |
| CN105320919A (en) | Human eye positioning method and apparatus | |
| JP3335542B2 (en) | Method and apparatus for collating handwritten characters | |
| Harfiya et al. | Offline signature verification based on pyramid histogram of oriented gradient features | |
| KR101972231B1 (en) | Apparatus and method of signature authentication using deep learning | |
| Tolosana et al. | Optimal feature selection and inter-operability compensation for on-line biometric signature authentication | |
| Boutellaa et al. | Improving online signature verification by user-specific likelihood ratio score normalization | |
| Shrivastava et al. | Fingerprint identification using feature extraction: A survey | |
| CN115171223A (en) | Online signature identification method and system based on unconstrained dynamic time programming | |
| CN110390268B (en) | Three-dimensional palmprint recognition method based on geometric characteristics and direction characteristics | |
| Wang et al. | On-line signature verification using segment-to-segment graph matching |
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 |