CN105718918A - Video based authentication method for signature in air - Google Patents
Video based authentication method for signature in air Download PDFInfo
- Publication number
- CN105718918A CN105718918A CN201610059700.7A CN201610059700A CN105718918A CN 105718918 A CN105718918 A CN 105718918A CN 201610059700 A CN201610059700 A CN 201610059700A CN 105718918 A CN105718918 A CN 105718918A
- Authority
- CN
- China
- Prior art keywords
- signature
- registration
- authentication
- needing
- pursuit path
- 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.)
- Granted
Links
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
- G06V40/30—Writer recognition; Reading and verifying 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
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention provides a video based authentication method for a signature in the air. The method comprises that features of the signature to be authenticated are extracted to obtain a track feature of the signature to be authenticated, and features of a registered signature are extracted to obtain a track feature of the registered signature; and the track feature of the signature to be authenticated matches the track feature of the registered signature to realize authentication for the signature. The video based authentication method can be used to effectively improve the identification precision of the signature, and further improve the accuracy and efficacy of signature authentication.
Description
Technical field
The present invention relates to image processing and analysis technical field, more particularly, it relates to a kind of aerial signature authentication method based on video.
Background technology
In recent years, fast development and extensive use along with biometrics identification technology, biometric authentication technology is proposed more higher requirement by user, biological characteristic authentication is made to occur in that some new development trends, such as higher antifalsification, higher accuracy of identification, and higher user's acceptance.
Although on-line signature certification obtains quickly development and application, but also has some shortcomings, it is mainly reflected in the following aspects:
(1) stroke and the number of words signed are limited, are therefore difficult to accurate assurance when choosing and having representational feature；
(2) when the different time, changed by different writing implements and psychology physiological situation be all likely to result in before and after signature inconsistent, therefore compared with other biological characteristic such as fingerprint etc., less stable, accuracy rate is on the low side；
(3) the form domination signed so that some static natures of signature are easily stolen to be taken and forge, and therefore antifalsification is under some influence；
(4) it is also made to there is the problem that application scenario is partially narrow based on the condition that depends on of board.
These problems above-mentioned are all the places that on-line signature certification faces improvement.
Summary of the invention
It is an object of the invention to overcome shortcoming of the prior art with not enough, it is provided that a kind of aerial signature authentication method based on video, the method can be effectively improved the accuracy of identification of signature, thus improving accuracy rate and the effectiveness of signature authentication.
In order to achieve the above object, the technical scheme is that: a kind of aerial signature authentication method based on video, it is characterized in that: the signature of the signature and registration that need certification is carried out feature extraction, respectively obtain the track characteristic needing authentication signature and the track characteristic of registration signature；Again the track characteristic of the track characteristic and registration signature that need authentication signature is mated the certification realizing signature；
Wherein, the signature of registration is carried out feature extraction to refer to: by gathering the signature video of registration, and the finger tip of registration in video is detected and followed the tracks of, obtain the signature track produced when finger is aloft write；Then the signature track of registration is carried out pretreatment, and extract the signature track characteristic of registration；
The signature needing certification is carried out feature extraction refer to: by gathering the signature video needing certification, and the finger tip needing certification in video is detected and followed the tracks of, obtain the signature track produced when finger is aloft write；Then the signature track needing certification is carried out pretreatment, and extracts the signature track characteristic needing certification.
In such scheme, the signature of registration and need the signature of certification all to adopt identical processing method, then can ensure the concordance of coupling, thus being effectively improved the accuracy of identification of signature.
The finger tip of registration in video is detected and follows the tracks of and the finger tip needing certification in video is detected and followed the tracks of, all adopts the TLD algorithm of improvement, respectively obtain the pursuit path of registration signature and need the pursuit path of authentication signature.
The TLD algorithm of described improvement comprises the following steps:
The first step, reads video file or directly reads photographic head real-time pictures, and frame goes out three finger tip kneading portion, completes target frame and initializes；
Second step, utilize and improve TLD track algorithm tracking finger tip target: joining day contextual information during following the tracks of, ratio according to image size with target frame size is adaptively adjusted the size of region of search, and adopts the detection of TLD algorithm in region of search and follow the tracks of finger tip target；
3rd step, sets frame number threshold value T3, accumulative region of search is not detected by the frame number D of finger tip target, by frame number D and frame number threshold value T3Relatively judge that whether tracking is failed: if D is > T3, then failure is followed the tracks of, it is proposed to warn and return the first step and again follow the tracks of；Otherwise, preserve the track of signature, respectively obtain the pursuit path of registration signature and need the pursuit path of authentication signature.
As it has been described above, the TLD algorithm improved is based on the TLD track algorithm of time context, it suitably reduces detection, following range to avoid unnecessary detection and the impact of part background, and then improves the tracking velocity of algorithm and the accuracy rate of algorithm.Great many of experiments displays that, by the TLD algorithm improved, big in intensity of illumination, environment is simple, when finger tip rapid movement, for the good effect that the upper and lower, left and right of finger tip all took time inclined, has strong robustness.
The signature track registered is carried out pretreatment comprise the following steps:
The first step, setting speed threshold value T1, calculate the speed V obtaining head and the tail tracing point in the pursuit path of registration signature1, by speed V1With threshold speed T1Relatively judge whether this tracing point is mistake tracing point, obtain the pursuit path one of registration signature: if V1<T1, then judge that this tracing point as mistake tracing point and is deleted；Otherwise retain this tracing point；
Second step, setpoint distance threshold value T2, calculate the distance L of each two tracing point in the pursuit path one of registration signature1, will apart from L1With distance threshold T2Relatively judge whether to follow the tracks of exception or loss of data: if L1>T2, then judge that the pursuit path one of registration signature follows the tracks of exception or loss of data, the result of front and back two point interpolation computing substituted this erroneous point, obtains the pursuit path two of registration signature；Otherwise, it is judged that the pursuit path one of registration signature is followed the tracks of normally, obtains the pursuit path two of registration signature；
3rd step, carries out dimension normalization to the position coordinates of the pursuit path two of registration signature, obtains the track characteristic of registration signature.
The signature track needing certification is carried out pretreatment comprise the following steps:
The first step, setting speed threshold value T1, calculate the speed V obtaining head and the tail tracing point in the pursuit path needing authentication signature2, by speed V2With threshold speed T1Relatively judge whether this tracing point is mistake tracing point, obtain needing the pursuit path one of authentication signature: if V2<T1, then judge that this tracing point as mistake tracing point and is deleted；Otherwise retain this tracing point；
Second step, setpoint distance threshold value T2, calculate the distance L of each two tracing point in the pursuit path one needing authentication signature2, will apart from L2With distance threshold T2Relatively judge whether to follow the tracks of exception or loss of data: if L2>T2, then judge to need the pursuit path one of authentication signature to follow the tracks of exception or loss of data, the result of front and back two point interpolation computing substituted this erroneous point, obtains needing the pursuit path two of authentication signature；Otherwise, it is judged that need the pursuit path one of authentication signature to follow the tracks of normally, obtain needing the pursuit path two of authentication signature；
3rd step, carries out dimension normalization to the position coordinates of the pursuit path two needing authentication signature, obtains needing the track characteristic of authentication signature.
In reality is signed, due to problems such as the TLD algorithm improved and SRTs, several track data points before just having started and having terminated overlap in a large number and shake, and have a strong impact on recognition effect, therefore adopt threshold speed T by the above-mentioned first step1Removing the data point of these mistakes, remove head and the tail coincidence tracing point and fine jitter tracing point, the speed that head and the tail tracing point is obtained is less than threshold speed T1Tracing point remove, thus obtaining the pursuit path one of registration signature and needing the pursuit path one of authentication signature.
When track algorithm runs, it is possible to because the reason such as hardware and video frame losing, cause part data point to be lost or abnormal, cause error in data.Second step of the present invention adopts distance threshold algorithm, calculates the distance between each two tracing point, arranges a bigger threshold value, it is judged that the erroneous point that is likely to occur also obtains the pursuit path two of registration signature with interpolation algorithm correction and needs the pursuit path two of authentication signature.
Consider that the initial of signature and extreme point are different, affect recognition effect, require over the 3rd step position coordinates to the pursuit path two of registration signature to be normalized and the position coordinates of the pursuit path two needing authentication signature is normalized, by registering signature size and needing authentication signature scaled as same size, to eliminate the impact that signatures match precision is caused by size.
The track characteristic that the track characteristic and registration that need authentication signature are signed mates the certification realizing signature and refers to: by the combination of dynamic time warping and fast fourier transform algorithm, the track characteristic of the track characteristic and registration signature that need authentication signature is mated, it is achieved the certification of signature.
Different from on-line signature, the new authentication method that the present invention proposes carries out in the motor process signed in finger tip pinch chalaza, do not have based on pressure utilizable in board signature process and the information such as pen state of rising and falling, but contain longer handwriting trace, and other characteristic informations such as different presentation direction, therefore contain the characteristic information of more horn of plenty.For these features, aerial signature is authenticated by the present invention based on dynamic time programming (DTW) algorithm and fast Fourier transform (FFT) algorithm.
The described combination by dynamic time warping and fast fourier transform algorithm, mates the track characteristic of the track characteristic and registration signature that need authentication signature, it is achieved the certification of signature, comprises the following steps:
The first step, chooses the track characteristic of N number of registration signature and calculates average length, according to the average length of the track characteristic of N number of registration signature, the track characteristic of N number of registration signature is interpolated computing, obtains mean trajectory feature as matching template；Wherein, 2 < N≤registration total sample number；
Second step, calculates DTW distance and the FFT distance of track characteristic and the matching template of N number of registration signature respectively, and calculates the meansigma methods one and the variance one that obtain DTW distance, and the meansigma methods two of FFT distance and variance two；Then associating gaussian probability model is set up according to meansigma methods one, variance one, meansigma methods two and variance two；
3rd step, calculates the arbitrary track characteristic of authentication signature and the DTW distance of matching template and FFT distance of needing, and then substitutes into associating gaussian probability model calculating matching probability；Matching probability is judged to actual signature more than the track characteristic needing authentication signature setting threshold value, is otherwise judged to forge a signature.
Wherein, DTW distance refers to the similarity between the coordinate using calculated two signatures of DTW algorithm, and FFT distance refers to the co sinus vector included angle value of 32 dimension Fourier coefficients before the X of two signatures, Y coordinate are obtained after carrying out FFT.
Signature Authentication System is it is generally required to obtain matching template by registering signature, because the length of signature is often different, it is necessary to be interpolated computing and obtain the coordinate sequence of equal length, be averaged one matching template of acquisition to each coordinate points.The template obtained by this average calculating operation can well avoid certain registration signature and other registration signature excessive harmful effects caused of gap.
Because needing the signature of certification to also tend to different from the length of matching template, the distance of two signatures can not be made directly calculating.DTW algorithm can be found an optimal trajectory and mate the corresponding point of two different sequences of length, calculates the Euclidean distance of summation corresponding point as DTW distance.Time-domain signal can be converted to frequency-region signal by Fourier transformation, takes front 32 coefficients and can well express former signature, and the co sinus vector included angle of 32 dimensional feature vectors of two signatures can as the distance metric of FFT.DTW and fft algorithm from the distance between time domain and two angle calculation signatures of frequency domain, obtain mutually supplementing to improve algorithm performance respectively over time and space.
Compared with prior art, the invention have the advantages that and beneficial effect: the present invention can be effectively improved the accuracy of identification of signature based on the aerial signature authentication method of video, thus improving accuracy rate and the effectiveness of signature authentication.
Accompanying drawing explanation
Fig. 1 is the present invention FB(flow block) based on the aerial signature authentication method of video；
Fig. 2 (1)-Fig. 2 (8) is the design sketch followed the tracks of after the TLD algorithm improved；
Fig. 3 is the trajectory diagram that finger tip follows the tracks of result；
Fig. 4 is the trajectory diagram of tape error tracing point；
Detailed description of the invention
Below in conjunction with accompanying drawing, the present invention is described in further detail with detailed description of the invention.
Embodiment
As it is shown in figure 1, a kind of aerial signature authentication method based on video of the present invention, it is that the signature to the signature and registration that need certification carries out feature extraction, respectively obtains the track characteristic needing authentication signature and the track characteristic of registration signature；Again the track characteristic of the track characteristic and registration signature that need authentication signature is mated the certification realizing signature；
Wherein, the signature of registration is carried out feature extraction to refer to: by gathering the signature video of registration, and the finger tip of registration in video is detected and followed the tracks of, obtain the signature track produced when finger is aloft write；Then the signature track of registration is carried out pretreatment, and extract the signature track characteristic of registration；
The signature needing certification is carried out feature extraction refer to: by gathering the signature video needing certification, and the finger tip needing certification in video is detected and followed the tracks of, obtain the signature track produced when finger is aloft write；Then the signature track needing certification is carried out pretreatment, and extracts the signature track characteristic needing certification.
The above-mentioned finger tip to registering in video detects and follows the tracks of and the finger tip needing certification in video is detected and followed the tracks of, and all adopts the TLD algorithm of improvement, respectively obtains the pursuit path of registration signature and needs the pursuit path of authentication signature.Wherein, the TLD algorithm of improvement comprises the following steps:
The first step, reads video file or directly reads photographic head real-time pictures, and frame goes out three finger tip kneading portion, completes target frame and initializes；
Second step, utilize and improve TLD track algorithm tracking finger tip target: joining day contextual information during following the tracks of, ratio according to image size with target frame size is adaptively adjusted the size of region of search, and adopts the detection of TLD algorithm in region of search and follow the tracks of finger tip target；
3rd step, sets frame number threshold value T3, accumulative region of search is not detected by the frame number D of finger tip target, by frame number D and frame number threshold value T3Relatively judge that whether tracking is failed: if D is > T3, then failure is followed the tracks of, it is proposed to warn and return the first step and again follow the tracks of；Otherwise, preserve the track of signature, respectively obtain the pursuit path of registration signature and need the pursuit path of authentication signature.
The TLD algorithm improved is based on the TLD track algorithm of time context, and it suitably reduces detection, following range to avoid unnecessary detection and the impact of part background, and then improves the tracking velocity of algorithm and the accuracy rate of algorithm.Shown in the design sketch such as Fig. 2 (1)-Fig. 2 (2) that the present invention follows the tracks of after adopting the TLD algorithm of improvement, and the trajectory diagram of finger tip tracking result is as shown in Figure 3.Great many of experiments displays that, by the TLD algorithm improved, big in intensity of illumination, environment is simple, when finger tip rapid movement, for the good effect that the upper and lower, left and right of finger tip all took time inclined, has strong robustness.
The signature track of registration is carried out pretreatment and comprises the following steps by the present invention:
The first step, setting speed threshold value T1, calculate the speed V obtaining head and the tail tracing point in the pursuit path of registration signature1, by speed V1With threshold speed T1Relatively judge whether this tracing point is mistake tracing point, obtain the pursuit path one of registration signature: if V1<T1, then judge that this tracing point as mistake tracing point and is deleted；Otherwise retain this tracing point；
Second step, setpoint distance threshold value T2, calculate the distance L of each two tracing point in the pursuit path one of registration signature1, will apart from L1With distance threshold T2Relatively judge whether to follow the tracks of exception or loss of data: if L1>T2, then judge that the pursuit path one of registration signature follows the tracks of exception or loss of data, the result of front and back two point interpolation computing substituted this erroneous point, obtains the pursuit path two of registration signature；Otherwise, it is judged that the pursuit path one of registration signature is followed the tracks of normally, obtains the pursuit path two of registration signature；
3rd step, carries out dimension normalization to the position coordinates of the pursuit path two of registration signature, obtains the track characteristic of registration signature.
The signature track needing certification is carried out pretreatment comprise the following steps:
The first step, setting speed threshold value T1, calculate the speed V obtaining head and the tail tracing point in the pursuit path needing authentication signature2, by speed V2With threshold speed T1Relatively judge whether this tracing point is mistake tracing point, obtain needing the pursuit path one of authentication signature: if V2<T1, then judge that this tracing point as mistake tracing point and is deleted；Otherwise retain this tracing point；
Second step, setpoint distance threshold value T2, calculate the distance L of each two tracing point in the pursuit path one needing authentication signature2, will apart from L2With distance threshold T2Relatively judge whether to follow the tracks of exception or loss of data: if L2>T2, then judge to need the pursuit path one of authentication signature to follow the tracks of exception or loss of data, the result of front and back two point interpolation computing substituted this erroneous point, obtains needing the pursuit path two of authentication signature；Otherwise, it is judged that need the pursuit path one of authentication signature to follow the tracks of normally, obtain needing the pursuit path two of authentication signature；
3rd step, carries out dimension normalization to the position coordinates of the pursuit path two needing authentication signature, obtains needing the track characteristic of authentication signature.
In reality is signed, due to problems such as the TLD algorithm improved and SRTs, several track data points before just having started and having terminated overlap in a large number and shake, have a strong impact on recognition effect (as shown in Figure 4, with the tracing point of overlapping tracing point or fine jitter in figure), therefore adopt threshold speed T by the above-mentioned first step1Removing the data point of these mistakes, remove head and the tail coincidence tracing point and fine jitter tracing point, the speed that head and the tail tracing point is obtained is less than threshold speed T1Tracing point remove, thus obtaining the pursuit path one of registration signature and needing the pursuit path one of authentication signature.
When track algorithm runs, it is possible to because the reason such as hardware and video frame losing, cause part data point to be lost or abnormal, cause error in data.Second step of the present invention adopts distance threshold algorithm, calculates the distance between each two tracing point, arranges a bigger threshold value, it is judged that the erroneous point that is likely to occur also obtains the pursuit path two of registration signature with interpolation algorithm correction and needs the pursuit path two of authentication signature.
Consider that the initial of signature and extreme point are different, affect recognition effect, require over the 3rd step position coordinates to the pursuit path two of registration signature to be normalized and the position coordinates of the pursuit path two needing authentication signature is normalized, by registering signature size and needing authentication signature scaled as same size, to eliminate the impact that signatures match precision is caused by size.
The track characteristic of present invention track characteristic and registration signature to needing authentication signature mates the certification realizing signature and refers to: by the combination of dynamic time warping and fast fourier transform algorithm, the track characteristic of the track characteristic and registration signature that need authentication signature is mated, it is achieved the certification of signature.
Different from on-line signature, the new authentication method that the present invention proposes carries out in the motor process signed in finger tip pinch chalaza, do not have based on pressure utilizable in board signature process and the information such as pen state of rising and falling, but contain longer handwriting trace, and other characteristic informations such as different presentation direction, therefore contain the characteristic information of more horn of plenty.For these features, aerial signature is authenticated by the present invention based on dynamic time programming (DTW) algorithm and fast Fourier transform (FFT) algorithm.
Above by the combination of dynamic time warping and fast fourier transform algorithm, the track characteristic of the track characteristic and registration signature that need authentication signature is mated, it is achieved the certification of signature, comprises the following steps:
The first step, chooses the track characteristic of five registration signatures and calculates average length, according to the average length of the track characteristic of five registration signatures, the track characteristic of five registration signatures is interpolated computing, obtains mean trajectory feature as matching template；
Second step, calculates DTW distances and the FFT distance of track characteristics and the matching template of five registration signatures respectively, and calculates the meansigma methods one and the variance one that obtain DTW distance, and the meansigma methods two of FFT distance and variance two；Then associating gaussian probability model is set up according to meansigma methods one, variance one, meansigma methods two and variance two；
3rd step, calculates the arbitrary track characteristic of authentication signature and the DTW distance of matching template and FFT distance of needing, and then substitutes into associating gaussian probability model calculating matching probability；The matching probability track characteristic needing authentication signature more than 50% is judged to actual signature, is otherwise judged to forge a signature.
Wherein, DTW distance refers to the similarity between the coordinate using calculated two signatures of DTW algorithm, and FFT distance refers to the co sinus vector included angle value of 32 dimension Fourier coefficients before the X of two signatures, Y coordinate are obtained after carrying out FFT.
Signature Authentication System is it is generally required to obtain matching template by registering signature, because the length of signature is often different, it is necessary to be interpolated computing and obtain the coordinate sequence of equal length, be averaged one matching template of acquisition to each coordinate points.The template obtained by this average calculating operation can well avoid certain registration signature and other registration signature excessive harmful effects caused of gap.
Because needing the signature of certification to also tend to different from the length of matching template, the distance of two signatures can not be made directly calculating.DTW algorithm can be found an optimal trajectory and mate the corresponding point of two different sequences of length, calculates the Euclidean distance of summation corresponding point as DTW distance.Time-domain signal can be converted to frequency-region signal by Fourier transformation, takes front 32 coefficients and can well express former signature, and the co sinus vector included angle of 32 dimensional feature vectors of two signatures can as the distance metric of FFT.DTW and fft algorithm from the distance between time domain and two angle calculation signatures of frequency domain, obtain mutually supplementing to improve algorithm performance respectively over time and space.
The present invention can be applicable to mobile phone, palm handheld terminal (such as palm PC), the processor with video camera and the computer with video camera etc. based on the aerial signature authentication method of video.When being applied on mobile phone, can gathering the signature video of registration by the photographic head of mobile phone and need the signature video of certification, user can apply the APP software installing signature authentication method of the present invention on mobile phone and realize the certification of signature.
Above-described embodiment is the present invention preferably embodiment; but embodiments of the present invention are also not restricted to the described embodiments; the change made under other any spirit without departing from the present invention and principle, modification, replacement, combination, simplification; all should be the substitute mode of equivalence, be included within protection scope of the present invention.
Claims (7)
1. the aerial signature authentication method based on video, it is characterised in that: the signature of the signature and registration that need certification is carried out feature extraction, respectively obtains the track characteristic needing authentication signature and the track characteristic of registration signature；Again the track characteristic of the track characteristic and registration signature that need authentication signature is mated the certification realizing signature；
Wherein, the signature of registration is carried out feature extraction to refer to: by gathering the signature video of registration, and the finger tip of registration in video is detected and followed the tracks of, obtain the signature track produced when finger is aloft write；Then the signature track of registration is carried out pretreatment, and extract the signature track characteristic of registration；
The signature needing certification is carried out feature extraction refer to: by gathering the signature video needing certification, and the finger tip needing certification in video is detected and followed the tracks of, obtain the signature track produced when finger is aloft write；Then the signature track needing certification is carried out pretreatment, and extracts the signature track characteristic needing certification.
2. the aerial signature authentication method based on video according to claim 1, it is characterized in that: the finger tip of registration in video is detected and follows the tracks of and the finger tip needing certification in video is detected and followed the tracks of, all adopt the TLD algorithm of improvement, respectively obtain the pursuit path of registration signature and need the pursuit path of authentication signature.
3. the aerial signature authentication method based on video according to claim 2, it is characterised in that: the TLD algorithm of described improvement comprises the following steps:
The first step, reads video file or directly reads photographic head real-time pictures, and frame goes out three finger tip kneading portion, completes target frame and initializes；
Second step, utilize and improve TLD track algorithm tracking finger tip target: joining day contextual information during following the tracks of, ratio according to image size with target frame size is adaptively adjusted the size of region of search, and adopts the detection of TLD algorithm in region of search and follow the tracks of finger tip target；
3rd step, sets frame number threshold value T3, accumulative region of search is not detected by the frame number D of finger tip target, by frame number D and frame number threshold value T3Relatively judge that whether tracking is failed: if D is > T3, then failure is followed the tracks of, it is proposed to warn and return the first step and again follow the tracks of；Otherwise, preserve the track of signature, respectively obtain the pursuit path of registration signature and need the pursuit path of authentication signature.
4. the aerial signature authentication method based on video according to claim 3, it is characterised in that: the signature track registered is carried out pretreatment and comprises the following steps:
The first step, setting speed threshold value T1, calculate the speed V obtaining head and the tail tracing point in the pursuit path of registration signature1, by speed V1With threshold speed T1Relatively judge whether this tracing point is mistake tracing point, obtain the pursuit path one of registration signature: if V1<T1, then judge that this tracing point as mistake tracing point and is deleted；Otherwise retain this tracing point；
Second step, setpoint distance threshold value T2, calculate the distance L of each two tracing point in the pursuit path one of registration signature1, will apart from L1With distance threshold T2Relatively judge whether to follow the tracks of exception or loss of data: if L1>T2, then judge that the pursuit path one of registration signature follows the tracks of exception or loss of data, the result of front and back two point interpolation computing substituted this erroneous point, obtains the pursuit path two of registration signature；Otherwise, it is judged that the pursuit path one of registration signature is followed the tracks of normally, obtains the pursuit path two of registration signature；
3rd step, carries out dimension normalization to the position coordinates of the pursuit path two of registration signature, obtains the track characteristic of registration signature.
5. the aerial signature authentication method based on video according to claim 3, it is characterised in that: the signature track needing certification is carried out pretreatment and comprises the following steps:
The first step, setting speed threshold value T1, calculate the speed V obtaining head and the tail tracing point in the pursuit path needing authentication signature2, by speed V2With threshold speed T1Relatively judge whether this tracing point is mistake tracing point, obtain needing the pursuit path one of authentication signature: if V2<T1, then judge that this tracing point as mistake tracing point and is deleted；Otherwise retain this tracing point；
Second step, setpoint distance threshold value T2, calculate the distance L of each two tracing point in the pursuit path one needing authentication signature2, will apart from L2With distance threshold T2Relatively judge whether to follow the tracks of exception or loss of data: if L2>T2, then judge to need the pursuit path one of authentication signature to follow the tracks of exception or loss of data, the result of front and back two point interpolation computing substituted this erroneous point, obtains needing the pursuit path two of authentication signature；Otherwise, it is judged that need the pursuit path one of authentication signature to follow the tracks of normally, obtain needing the pursuit path two of authentication signature；
3rd step, carries out dimension normalization to the position coordinates of the pursuit path two needing authentication signature, obtains needing the track characteristic of authentication signature.
6. the aerial signature authentication method based on video according to claim 1, it is characterized in that: the track characteristic that the track characteristic and registration that need authentication signature are signed mates the certification realizing signature and refers to: by the combination of dynamic time warping and fast fourier transform algorithm, the track characteristic of the track characteristic and registration signature that need authentication signature is mated, it is achieved the certification of signature.
7. the aerial signature authentication method based on video according to claim 6, it is characterized in that: the described combination by dynamic time warping and fast fourier transform algorithm, the track characteristic of the track characteristic and registration signature that need authentication signature is mated, realize the certification of signature, comprise the following steps:
The first step, chooses the track characteristic of N number of registration signature and calculates average length, according to the average length of the track characteristic of N number of registration signature, the track characteristic of N number of registration signature is interpolated computing, obtains mean trajectory feature as matching template；Wherein, 2 < N < registration total sample number；
Second step, calculates DTW distance and the FFT distance of track characteristic and the matching template of N number of registration signature respectively, and calculates the meansigma methods one and the variance one that obtain DTW distance, and the meansigma methods two of FFT distance and variance two；Then associating gaussian probability model is set up according to meansigma methods one, variance one, meansigma methods two and variance two；
3rd step, calculates the arbitrary track characteristic of authentication signature and the DTW distance of matching template and FFT distance of needing, and then substitutes into associating gaussian probability model calculating matching probability；Matching probability is judged to actual signature more than the track characteristic needing authentication signature setting threshold value, is otherwise judged to forge a signature.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610059700.7A CN105718918B (en) | 2016-01-28 | 2016-01-28 | Aerial signature authentication method based on video |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610059700.7A CN105718918B (en) | 2016-01-28 | 2016-01-28 | Aerial signature authentication method based on video |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105718918A true CN105718918A (en) | 2016-06-29 |
| CN105718918B CN105718918B (en) | 2019-07-16 |
Family
ID=56154263
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610059700.7A Active CN105718918B (en) | 2016-01-28 | 2016-01-28 | Aerial signature authentication method based on video |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105718918B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107944336A (en) * | 2017-10-11 | 2018-04-20 | 中国科学院自动化研究所 | Handwriting signature verification system based on cloud computing |
| CN109409316A (en) * | 2018-11-07 | 2019-03-01 | 极鱼（北京）科技有限公司 | Aerial endorsement method and device |
| CN109448189A (en) * | 2018-10-29 | 2019-03-08 | 极鱼（北京）科技有限公司 | Self-service bank's management method, self-service bank and self-service bank's system |
| CN113158887A (en) * | 2021-04-19 | 2021-07-23 | 福建捷宇电脑科技有限公司 | Electronic signature authentication method and equipment for improving identification accuracy of electronic signature |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573451A (en) * | 2013-10-29 | 2015-04-29 | 深圳市腾讯计算机系统有限公司 | User verification method and device and electronic device |
| US20150286872A1 (en) * | 2011-06-20 | 2015-10-08 | University Of Southern California | Visual tracking in video images in unconstrained environments by exploiting on-the-fly contxt using supporters and distracters |
| CN104992192A (en) * | 2015-05-12 | 2015-10-21 | 浙江工商大学 | Visual motion tracking telekinetic handwriting system |
-
2016
- 2016-01-28 CN CN201610059700.7A patent/CN105718918B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150286872A1 (en) * | 2011-06-20 | 2015-10-08 | University Of Southern California | Visual tracking in video images in unconstrained environments by exploiting on-the-fly contxt using supporters and distracters |
| CN104573451A (en) * | 2013-10-29 | 2015-04-29 | 深圳市腾讯计算机系统有限公司 | User verification method and device and electronic device |
| CN104992192A (en) * | 2015-05-12 | 2015-10-21 | 浙江工商大学 | Visual motion tracking telekinetic handwriting system |
Non-Patent Citations (2)
| Title |
|---|
| 周瑜等: "形状匹配方法研究与展望", 《自动化学报》 * |
| 张毅等: ""一种改进的TLD动态手势跟踪算法"", 《机器人》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107944336A (en) * | 2017-10-11 | 2018-04-20 | 中国科学院自动化研究所 | Handwriting signature verification system based on cloud computing |
| CN109448189A (en) * | 2018-10-29 | 2019-03-08 | 极鱼（北京）科技有限公司 | Self-service bank's management method, self-service bank and self-service bank's system |
| CN109409316A (en) * | 2018-11-07 | 2019-03-01 | 极鱼（北京）科技有限公司 | Aerial endorsement method and device |
| CN109409316B (en) * | 2018-11-07 | 2022-04-01 | 极鱼（北京）科技有限公司 | Over-the-air signature method and device |
| CN113158887A (en) * | 2021-04-19 | 2021-07-23 | 福建捷宇电脑科技有限公司 | Electronic signature authentication method and equipment for improving identification accuracy of electronic signature |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105718918B (en) | 2019-07-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11151363B2 (en) | Expression recognition method, apparatus, electronic device, and storage medium | |
| US10956719B2 (en) | Depth image based face anti-spoofing | |
| WO2017133605A1 (en) | Method and device for facial tracking and smart terminal | |
| US8644560B2 (en) | Image processing apparatus and method, and program | |
| WO2019071664A1 (en) | Human face recognition method and apparatus combined with depth information, and storage medium | |
| CN105718918A (en) | Video based authentication method for signature in air | |
| Das et al. | Recent advances in biometric technology for mobile devices | |
| Malassiotis et al. | Personal authentication using 3-D finger geometry | |
| US8243993B2 (en) | Method for moving object detection and hand gesture control method based on the method for moving object detection | |
| Kumar et al. | 3D text segmentation and recognition using leap motion | |
| CN111079791A (en) | Face recognition method, face recognition device and computer-readable storage medium | |
| US7760916B2 (en) | Registration apparatus, collation apparatus, image correction method and program | |
| CN110287918B (en) | Living body identification method and related product | |
| JP2008250999A (en) | Object tracing method, object tracing device and object tracing program | |
| US20200302218A1 (en) | Fast curve matching for tattoo recognition and identification | |
| CN109937434A (en) | Image processing method, device, terminal and storage medium | |
| KR20140055172A (en) | Apparatus and method for recognizing posture and gesture based depth data | |
| JP2006331271A (en) | Representative image extraction apparatus and program | |
| CN108255352B (en) | Multi-touch implementation method and system in projection interaction system | |
| Shariat et al. | A new adaptive segmental matching measure for human activity recognition | |
| CN113469220A (en) | Track identification method and related device and equipment | |
| WO2024045454A1 (en) | Target identification method, storage medium and device | |
| KR101001184B1 (en) | Iterative 3D head pose estimation method using a face normal vector | |
| CN115171223A (en) | Online signature identification method and system based on unconstrained dynamic time programming | |
| US20030123734A1 (en) | Methods and apparatus for object recognition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |