CN106951832B - Verification method and device based on handwritten character recognition - Google Patents
Verification method and device based on handwritten character recognition Download PDFInfo
- Publication number
- CN106951832B CN106951832B CN201710113449.2A CN201710113449A CN106951832B CN 106951832 B CN106951832 B CN 106951832B CN 201710113449 A CN201710113449 A CN 201710113449A CN 106951832 B CN106951832 B CN 106951832B
- Authority
- CN
- China
- Prior art keywords
- character
- handwritten
- information
- verification
- module
- 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 OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/32—Digital ink
- G06V30/333—Preprocessing; Feature extraction
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; 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
- G06F18/2163—Partitioning the feature space
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/32—Digital ink
- G06V30/36—Matching; Classification
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Character Discrimination (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a verification method and a device based on handwritten character recognition, wherein the method comprises the following steps: s1: extracting reference characters from an alternative character library through a random algorithm, and generating a picture as reference verification information; s2: acquiring handwritten characters in a handwriting area, and generating a picture as handwriting verification information; s3: judging whether the handwritten character of the handwritten authentication information is consistent with the reference character of the reference authentication information through a font recognition algorithm, if so, executing step S4, and if not, failing the authentication; s4: extracting the characteristics of the handwritten verification information to obtain corresponding characteristic vectors; s5: and judging whether the feature vector of the handwriting verification information is consistent with the feature vector in the template library, and if not, failing the verification. The invention distinguishes human beings and computers by obtaining the handwritten character in the designated area, thereby preventing the influence caused by the cracking of the verification code by a malicious program; the safety of the user using the internet is improved.
Description
Technical Field
The invention belongs to the technical field of image recognition, and particularly relates to a verification method and device based on handwritten character recognition.
Background
The Turing test (English: computer Automated Public piping test to tell Computers and Humans Apart, abbreviated as CAPTCHA) for fully automatically distinguishing Computers and Humans is commonly called a verification code, and is a Public full-automatic program for distinguishing Computers and Humans from users. In a CAPTCHA test, a computer acting as a server automatically generates a question to be solved by a user. This question can be generated and evaluated by a computer, but must only be solved by a human. Since the computer is unable to solve the CAPTCHA's problem, the user who answers the question may be considered a human.
Existing verification codes can be classified into 3 types: character verification code, picture verification code and voice verification code. The verification principle of the existing verification code technology mainly comprises the following steps: the method comprises the steps that a group of character strings are randomly generated through a server end and serve as information to be verified, the random character strings are converted into information (such as pictures or voice) which can be easily recognized by human beings but cannot be recognized by a computer as reference verification codes through program processing as much as possible, the reference verification codes are displayed to a user through a browser end for recognition, the user submits the verification information to the server end through a mouse or keyboard event (such as mouse clicking and keyboard input) through a mouse or keyboard event after being recognized by the user, the server end distinguishes whether the user is a person or a computer through judging whether the submitted verification information is consistent with the information to be verified, if the submitted verification information is consistent with the information to be verified, the user is judged to be a person, and if the submitted verification information is not consistent with the submitted verification information, the user is judged to be a computer.
Disadvantages or shortcomings: in the existing verification code technology, the operation of submitting verification information by a user is through a mouse or keyboard event, and the type of submitting the verification information is characters; fig. 1 is a schematic diagram of a conventional verification code technology. In the existing verification code technology, a method for enabling a user to verify a verification code by using a handwritten character in a designated area does not exist, and the type of verification information submitted by the method is a picture of the handwritten character. In the existing verification code technology, because a computer can easily simulate a keyboard and a mouse event, and meanwhile, the technology for extracting the content of the picture or the voice verification code by the computer is mature, the computer can easily identify and extract the character content contained in the picture or the voice reference verification code through a visual recognition algorithm and a voice recognition algorithm, and simulate event behaviors of a user such as mouse click, keyboard input and the like to input the identified characters at a browser end for submission. Therefore, the way of human and computer authentication by text/image/voice verification codes has become rather unreliable.
In the existing verification code technology, the idea that the reference verification code is easy to recognize by human and not easy to recognize by a machine is mainly designed, and the difficulty of machine recognition is mainly increased by means of interfering with machine recognition, such as distorted characters, use of a miscellaneous background in a picture, increase of background noise in voice and the like. The use of such a mode also increases the difficulty of human recognition, and even the situation that the human cannot recognize the human is sometimes occurred, so that the user experience is reduced.
Disclosure of Invention
In order to overcome the defects of the prior art, one of the objectives of the present invention is to provide an authentication method based on handwritten character recognition, which can improve the security of the authentication recognition.
It is another object of the present invention to provide an authentication device based on handwritten character recognition, which can improve the security of authentication recognition.
One of the purposes of the invention is realized by adopting the following technical scheme:
a verification method based on handwritten character recognition comprises the following steps:
s1: extracting reference characters from an alternative character library through a random algorithm, and generating a picture as reference verification information;
s2: acquiring handwritten characters in a handwriting area, and generating a picture as handwriting verification information;
s3: judging whether the handwritten character of the handwritten verification information is consistent with the reference character through a font recognition algorithm, if so, executing step S4, and if not, failing the verification;
s4: extracting the characteristics of the handwritten verification information to obtain corresponding characteristic vectors;
s5: and judging whether the feature vector of the handwriting verification information is consistent with the feature vector in the template library, if so, the verification is successful, and if not, the verification is failed.
Preferably, the step S6 is executed after the step S5 is yes: and judging whether the similarity between the handwritten verification information and the submitted handwritten verification information is greater than a preset value, if so, failing the verification, and if not, succeeding the verification.
Preferably, step S21 is further included after step S2: and preprocessing the handwritten authentication information.
Preferably, step S21 specifically includes the following sub-steps:
s211: carrying out binarization processing on the handwritten verification information;
s212: performing character segmentation on the handwritten verification information through a seed connectivity algorithm to obtain character information;
s213: and carrying out normalization processing on the character information through centroid alignment and linear interpolation amplification.
Preferably, step S212 is followed by the steps of:
step S2120: judging whether the character features of the character information are in a preset range, if so, executing step S213, and if not, executing step S2121;
step S2121 judges whether the character information is a sticky character through pre-recognition, if so, it is segmented by a method of finding a valley point in the vertical projection diagram, and if not, step S213 is executed.
Preferably, step S4 specifically includes the following sub-steps:
s41: carrying out image segmentation on the character information, and segmenting the character information into square grid areas with preset quantity;
s42: and calculating the area density in each square, wherein the area density is the ratio of the number of points in each square to the total number of points of the character information.
Preferably, the preset number of the grid regions is 5 × 5 grid regions.
The second purpose of the invention is realized by adopting the following technical scheme:
an authentication apparatus based on handwritten character recognition, comprising the following modules:
a verification information generation module: the device comprises a random algorithm, a reference character database and a picture database, wherein the random algorithm is used for extracting reference characters from the alternative character database and generating the picture as reference verification information;
the handwritten information acquisition module: the handwriting verification system is used for acquiring handwritten characters in the handwriting area and generating pictures as handwriting verification information;
a character recognition module: the character recognition module is used for judging whether the handwritten character of the handwritten verification information is consistent with the reference character of the reference verification information through a font recognition algorithm, if so, the character extraction module is executed, and if not, the verification fails;
a feature extraction module: the system is used for extracting the characteristics of the handwritten verification information to obtain corresponding characteristic vectors;
a feature comparison module: and the method is used for judging whether the feature vector of the handwriting verification information is consistent with the feature vector in the template library or not, and if not, the verification fails.
Preferably, the similarity judging module is executed after the characteristic comparing module judges that the result is yes: and the method is used for judging whether the similarity between the handwritten verification information and the submitted handwritten verification information is greater than a preset value, if so, the verification fails, and if not, the verification succeeds.
Preferably, the handwriting information acquisition module further comprises a preprocessing module: for preprocessing the handwritten authentication information.
Preferably, the preprocessing module specifically includes the following sub-modules:
a binarization module: the handwriting verification information processing device is used for carrying out binarization processing on the handwriting verification information;
a character segmentation module: the system comprises a seed communication algorithm, a handwriting verification information acquisition module, a handwriting recognition module and a handwriting recognition module, wherein the seed communication algorithm is used for performing character segmentation on the handwriting verification information to obtain character information;
a normalization module: the method is used for normalizing the character information through centroid alignment and linear interpolation amplification.
Preferably, the character segmentation module further comprises the following modules:
a character characteristic judgment module: the device comprises a normalization module and a sticky character segmentation module, wherein the normalization module is used for judging whether the character characteristics of character information are in a preset range, if so, the normalization module is executed, and if not, the sticky character segmentation module is executed;
a sticky character segmentation module: and the normalization module is used for judging whether the character information is a sticky character or not through pre-recognition, if so, segmenting the sticky character through a method of finding valley points in the vertical projection image, and if not, executing the normalization module.
Compared with the prior art, the invention has the beneficial effects that:
the invention distinguishes human from computer by a method of writing characters in the designated area by the user, thereby effectively preventing the influence of the cracking of the verification code by a malicious program; the safety of the user using the internet is improved.
Drawings
FIG. 1 is a schematic diagram of a prior art captcha technique;
FIG. 2 is a flow chart of a handwritten character recognition based authentication method of the present invention;
fig. 3 is a block diagram of an authentication apparatus based on handwritten character recognition according to the present invention.
Detailed Description
The invention will be further described with reference to the accompanying drawings and the detailed description below:
as shown in fig. 2, the present invention provides a verification method based on handwritten character recognition, which includes the following steps:
before verification, a training identification template needs to be established firstly to carry out subsequent identification judgment; the learning verification is preprocessed, feature extracted and recognized, and a recognition model is built. Training: the method comprises the steps of extracting standard templates, namely a standard feature library (template library), from training set verification codes, wherein each handwritten character is provided with hundreds of standard templates, storing feature vectors of handwritten pictures of a training set into a file after preprocessing and feature extraction, indicating correct values of the handwritten pictures during training, and adding no template library into the handwritten pictures of the training set, which are found to be adhered by characters during segmentation, in order to avoid generating wrong standard templates.
S1: extracting reference characters from an alternative character library through a random algorithm, and generating a picture as reference verification information; the server randomly generates a group of character strings, stores the character strings as information to be verified, generates reference verification codes for the character strings and displays the reference verification codes to the user on a display terminal to enable the user to identify;
s2: acquiring handwritten characters in a handwriting area, and generating a picture as handwriting verification information; after acquiring verification information of handwritten characters in a designated area submitted by a user, the system performs feature extraction on the verification information of the handwritten pictures;
s21: preprocessing the handwritten verification information; step S21 specifically includes the following substeps: the preprocessing mainly comprises the steps of decoding, binaryzation, noise and interference removal, character segmentation, normalization and the like; the quality of preprocessing greatly affects the processing and identifying performance of a server on pictures, wherein the interference removal and character segmentation are particularly important; the preprocessing comprises a plurality of steps, the invention uses a plurality of main steps of decoding, binaryzation, noise and interference removal, character segmentation and normalization, can also add steps of smoothing processing and the like, and can also combine or singly use a plurality of or one step;
s211: carrying out binarization processing on the handwritten verification information; converting the gray value of the picture into 0 or 255 (namely black and white) by taking a certain threshold as a limit so as to be convenient for processing, and selecting a reasonable threshold to eliminate a lot of backgrounds and noises without damaging character strokes, wherein the binary threshold is obtained by analyzing the specific picture; removing interference points, wherein most of noise is removed after binarization, but a plurality of interference points exist, the interference points with the height of 1 pixel and 2 pixels can be removed by removing the interference points and the noise, and the subsequent processing can be better performed by removing the interference points;
s212: performing character segmentation on the handwritten verification information through a seed connectivity algorithm to obtain character information; step S212 is followed by the following steps:
step S2120: judging whether the character features of the character information are in a preset range, if so, executing step S213, and if not, executing step S2121; the character features mainly comprise the point number, the width and height features and the like of the character;
step S2121: judging whether the character information is a conglutinated character or not through pre-recognition, if so, segmenting the character information through a method of finding valley points in a vertical projection diagram, and if not, executing a step S213; the method comprises the steps of dividing handwritten picture information into single characters, obtaining a plurality of connecting lines by utilizing a seed filling algorithm, so that non-adhered characters can be divided, and further dividing the adhered characters, wherein the adhered characters are judged mainly according to the point number and the width and height characteristics of the characters, when the number of the non-adhered characters is larger than a certain threshold value, the non-adhered characters is preliminarily judged as character training, the threshold value is obtained through statistical analysis according to the picture characteristics, the preliminarily judged adhered characters are further judged by using a pre-recognition method in order to prevent judgment errors, and for the division of the adhered characters, a valley point finding method in vertical projection is adopted;
s213: carrying out normalization processing on the character information through centroid alignment and linear interpolation amplification; in order to solve the problems of position offset, different sizes and variable rotation of picture characters, the character information is normalized through centroid alignment and linear interpolation amplification to enable the characters to become a uniform rule so as to be convenient for matching;
s3: judging whether the handwritten character of the handwritten authentication information is consistent with the reference character of the reference authentication information through a font recognition algorithm, if so, executing step S4, and if not, failing the authentication; the font identification algorithm is the conventional technical means, can definitely identify whether the handwritten character is consistent with the reference character, and if not, the verification fails;
s4: extracting the characteristics of the handwritten verification information to obtain corresponding characteristic vectors; step S4 specifically includes the following substeps:
s41: carrying out image segmentation on the character information, and segmenting the character information into square grid areas with preset quantity; the grid areas with the preset number are 5-by-5 grid areas;
s42: calculating the area density in each square, wherein the area density is the ratio of the number of points in each square to the total number of points of the character information; extracting a feature vector with a certain dimension from a preprocessed character picture, thereby improving the storage capacity and the operation speed of character matching and recognition, wherein the character has a plurality of features, and the purpose of correct recognition can be achieved by selecting proper features; the distribution of the strokes of the space characters is reflected, so that the recognition and judgment among the characters are not carried out when training and recognition are carried out, but the stroke characteristics of the characters are continuously analyzed and then analyzed, so that corresponding judgment conditions are provided for a human-computer;
in the feature extraction link, the feature of extracting the regional density of the character is adopted, and other features such as smooth features and the like can be extracted to replace the feature; other characteristics can be added in the process of handwriting the character to judge whether the character is handwritten or not, for example, the characteristics of acquiring the handwriting order and the pressure feeling are added in the designated area to carry out auxiliary judgment.
S5: judging whether the feature vector of the handwriting verification information is consistent with the handwriting feature vector of the template library, if not, the verification fails;
the identification of the handwritten fonts is realized by using a softmax regression model, the softmax model can be used for distributing probability to different objects, and the softmax regression is mainly divided into two steps:
in the first step, in order to obtain the evidence that a given picture belongs to a certain type of characters, we add the weighted sum of the pixel values of the picture, if the pixel has strong evidence that the picture does not belong to the type, the corresponding weight is negative, otherwise, if the pixel has favorable evidence to support the picture to belong to the type, the corresponding weight is positive, and in addition, we add an extra offset to reject some irrelevant interference caused by the input, therefore,
wherein Wi,jRepresents a weight, xjRepresenting the feature vector, i.e. the region density in the present invention, biRepresents the offset of class i, j represents the pixel index of a given picture x for pixel summation, and then converts these evidences into a probability y using the softmax function: softmax (evidence); wherein Wi,jAnd offset biIs obtained through training, and is recognizedWaiting, the computer only needs to identify xjThis amount gives the corresponding evidenceiThus obtaining the corresponding probability y;
here softmax is an excitation function that converts the output of the linear function we define into the format we want, i.e. the probability distribution about the text class, so that, in the second step, given a picture, its goodness of fit for each training standard template can be converted into a probability value by the softmax function, which can be defined as softmax (x) normal (exp (x);
expand the right sub-formula of the equation to obtain
The probability that the character to be recognized belongs to which standard template is the largest indicates the character which is the best matched with that template, i.e. which character is distinguished. If yes, go to S6, if no, verify failure;
s6: and judging whether the similarity between the handwritten verification information and the submitted handwritten verification information is greater than a preset value, if so, failing the verification, and if not, succeeding the verification. The preset value is 100%; the submitted handwriting verification information is the verification information submitted to the computer terminal before the current handwriting verification information; the judgment principle of whether the similarity of the submitted handwriting verification information is 100% or not is as follows: because human handwriting input is different each time, all submitted handwritten verification information will not be 100% similar.
The step of the invention can only judge the consistency of the handwritten character and the reference verification information, and omit the step of comparing whether the similarity is 100% with the existing handwritten picture, wherein the step is used for judging the uniqueness of the handwritten picture, thereby preventing the same handwritten picture from being repeatedly submitted by a computer to replace a human handwritten picture to verify the consistency of the reference information.
When an instruction of a user for performing key operation is received, the system needs to judge and determine that the operation is initiated by a human or a machine, namely, the system enters human-machine verification, and the server judges whether submitted handwritten verification characters come from a real user or are generated by a computer through the steps of preprocessing, characteristic judgment, training, noise reduction, filtering, cutting and the like.
The method judges the effectiveness of a real user by the server-side handwritten character picture verification method, avoids the verification design by using the idea of interfering machine identification reference verification information and improving machine identification difficulty, so that the machine identification is interfered in the link of providing the reference verification information for the user without excessively increasing the technical means of complicated machine identification interference (distortion of characters, use of a miscellaneous point background in the picture and increase of background noise in voice), and the problem that the reference verification information is difficult to be identified by people due to the processing mode is avoided, thereby improving the user experience.
The invention uses machine learning algorithm to recognize and train a large number of handwritten Chinese and English characters, letters and number samples, establishes a recognition system, and realizes the rapid recognition of whether the handwritten character picture is from a real user or is generated by a machine through characteristic judgment.
As shown in fig. 3, the present invention provides a verification device based on handwritten character recognition, which includes the following modules:
a verification information generation module: the device comprises a random algorithm, a reference character database and a picture database, wherein the random algorithm is used for extracting reference characters from the alternative character database and generating the picture as reference verification information;
the handwritten information acquisition module: the handwriting verification system is used for acquiring handwritten characters in the handwriting area and generating pictures as handwriting verification information;
a character recognition module: the character recognition module is used for judging whether the handwritten character of the handwritten verification information is consistent with the reference character of the reference verification information through a font recognition algorithm, if so, the character extraction module is executed, and if not, the verification fails;
a preprocessing module: the system is used for preprocessing the handwriting verification information; the preprocessing module specifically comprises the following sub-modules;
a binarization module: the handwriting verification information processing device is used for carrying out binarization processing on the handwriting verification information;
a character segmentation module: the system comprises a seed communication algorithm, a handwriting verification information acquisition module, a handwriting recognition module and a handwriting recognition module, wherein the seed communication algorithm is used for performing character segmentation on the handwriting verification information to obtain character information; the character segmentation module is also followed by the following modules;
a character characteristic judgment module: the device comprises a normalization module and a sticky character segmentation module, wherein the normalization module is used for judging whether the character characteristics of character information are in a preset range, if so, the normalization module is executed, and if not, the sticky character segmentation module is executed;
a sticky character segmentation module: judging whether the character information is a sticky character or not through pre-recognition, if so, segmenting the sticky character information through a valley point finding method in a vertical projection image, and if not, executing a normalization module;
a normalization module: the character information is normalized through centroid alignment and linear interpolation amplification;
a feature extraction module: the system is used for extracting the characteristics of the handwritten verification information to obtain corresponding characteristic vectors;
a feature comparison module: the system is used for judging whether the feature vector of the handwritten verification information is consistent with the feature vector in the template library or not, if so, the similarity judging module is executed, and if not, the verification fails;
a similarity judging module: and the method is used for judging whether the similarity between the handwritten verification information and the submitted handwritten verification information is greater than a preset value, if so, the verification fails, and if not, the verification succeeds.
Various other modifications and changes may be made by those skilled in the art based on the above-described technical solutions and concepts, and all such modifications and changes should fall within the scope of the claims of the present invention.
Claims (9)
1. A verification method based on handwritten character recognition is characterized by comprising the following steps:
s1: extracting reference characters from an alternative character library through a random algorithm, and generating a picture as reference verification information;
s2: acquiring handwritten characters in a handwriting area, and generating a picture as handwriting verification information;
s3: judging whether the handwritten character of the handwritten authentication information is consistent with the reference character of the reference authentication information through a font recognition algorithm, if so, executing step S4, and if not, failing the authentication;
s4: extracting the characteristics of the handwritten verification information to obtain corresponding characteristic vectors;
s5: judging whether the feature vector of the handwriting verification information is consistent with the feature vector in the template library, if not, the verification fails;
after the step S5, executing the step S6: and judging whether the similarity between the handwritten verification information and the submitted handwritten verification information is greater than a preset value, if so, failing the verification, and if not, succeeding the verification.
2. The handwritten character recognition-based authentication method of claim 1, further comprising, after step S2, step S21: and preprocessing the handwritten authentication information.
3. The handwritten character recognition based authentication method according to claim 2, wherein the step S21 specifically comprises the following sub-steps:
s211: carrying out binarization processing on the handwritten verification information;
s212: performing character segmentation on the handwritten verification information through a seed connectivity algorithm to obtain character information;
s213: and carrying out normalization processing on the character information through centroid alignment and linear interpolation amplification.
4. The handwritten character recognition based authentication method according to claim 3, characterized in that step S212 is followed by the steps of:
step S2120: judging whether the character features of the character information are in a preset range, if so, executing step S213, and if not, executing step S2121;
step S2121: and judging whether the character information is a sticky character or not through pre-recognition, if so, segmenting the sticky character by a valley point finding method in the vertical projection drawing, and if not, executing the step S213.
5. The handwritten character recognition based authentication method according to claim 1, characterized in that step S4 specifically comprises the following sub-steps:
s41: carrying out image segmentation on the character information, and segmenting the character information into square grid areas with preset quantity;
s42: and calculating the area density in each square, wherein the area density is the ratio of the number of points in each square to the total number of points of the character information.
6. An authentication device based on handwritten character recognition, characterized by comprising the following modules:
a verification information generation module: the device comprises a random algorithm, a reference character database and a picture database, wherein the random algorithm is used for extracting reference characters from the alternative character database and generating the picture as reference verification information;
the handwritten information acquisition module: the handwriting verification system is used for acquiring handwritten characters in the handwriting area and generating pictures as handwriting verification information;
a character recognition module: the character recognition module is used for judging whether the handwritten character of the handwritten verification information is consistent with the reference character of the reference verification information through a font recognition algorithm, if so, the character extraction module is executed, and if not, the verification fails;
a feature extraction module: the system is used for extracting the characteristics of the handwritten verification information to obtain corresponding characteristic vectors;
a feature comparison module: the system is used for judging whether the feature vector of the handwriting verification information is consistent with the feature vector in the template library or not, and if not, the verification fails;
a similarity judging module: and the method is used for judging whether the similarity between the handwritten verification information and the submitted handwritten verification information is greater than a preset value, if so, the verification fails, and if not, the verification succeeds.
7. The handwritten character recognition based authentication device according to claim 6, further comprising a preprocessing module after the handwritten information acquisition module: for preprocessing the handwritten authentication information.
8. The handwritten character recognition based authentication device according to claim 7, wherein said preprocessing module comprises the following sub-modules:
a binarization module: the handwriting verification information processing device is used for carrying out binarization processing on the handwriting verification information;
a character segmentation module: the system comprises a seed communication algorithm, a handwriting verification information acquisition module, a handwriting recognition module and a handwriting recognition module, wherein the seed communication algorithm is used for performing character segmentation on the handwriting verification information to obtain character information;
a normalization module: the method is used for normalizing the character information through centroid alignment and linear interpolation amplification.
9. The handwritten character recognition based authentication device of claim 8, characterized in that said character segmentation module is followed by the following modules:
a character characteristic judgment module: the device comprises a normalization module and a sticky character segmentation module, wherein the normalization module is used for judging whether the character characteristics of character information are in a preset range, if so, the normalization module is executed, and if not, the sticky character segmentation module is executed;
a sticky character segmentation module: and judging whether the character information is a conglutinated character or not through pre-recognition, if so, segmenting the character information through a method of finding valley points in the vertical projection image, and if not, executing a normalization module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710113449.2A CN106951832B (en) | 2017-02-28 | 2017-02-28 | Verification method and device based on handwritten character recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710113449.2A CN106951832B (en) | 2017-02-28 | 2017-02-28 | Verification method and device based on handwritten character recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106951832A CN106951832A (en) | 2017-07-14 |
| CN106951832B true CN106951832B (en) | 2022-02-18 |
Family
ID=59467684
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710113449.2A Active CN106951832B (en) | 2017-02-28 | 2017-02-28 | Verification method and device based on handwritten character recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106951832B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107451559A (en) * | 2017-07-31 | 2017-12-08 | 邱宇轩 | Parkinson's people's handwriting automatic identifying method based on machine learning |
| CN107729804A (en) * | 2017-08-31 | 2018-02-23 | 广东数相智能科技有限公司 | A kind of people flow rate statistical method and device based on garment ornament |
| CN107894988A (en) * | 2017-09-29 | 2018-04-10 | 广东数相智能科技有限公司 | A kind of hunger sign in plant detection method, electronic equipment, storage medium and device |
| CN107909072A (en) * | 2017-09-29 | 2018-04-13 | 广东数相智能科技有限公司 | A kind of vegetation type recognition methods, electronic equipment, storage medium and device |
| CN108874756B (en) * | 2018-06-29 | 2022-05-20 | 广东智媒云图科技股份有限公司 | Verification code optimization method |
| CN109522898B (en) * | 2018-09-18 | 2024-10-15 | 平安科技（深圳）有限公司 | Handwritten sample picture labeling method and device, computer equipment and storage medium |
| CN109472234B (en) * | 2018-11-01 | 2021-07-20 | 北京爱知之星科技股份有限公司 | Intelligent recognition method for handwriting input |
| CN109583423A (en) * | 2018-12-18 | 2019-04-05 | 苏州大学 | A kind of method, apparatus and associated component of Handwritten Digit Recognition |
| CN109993202B (en) * | 2019-02-15 | 2023-08-22 | 广东智媒云图科技股份有限公司 | Line manuscript type graph similarity judging method, electronic equipment and storage medium |
| CN109948653B (en) * | 2019-02-15 | 2023-08-25 | 广东智媒云图科技股份有限公司 | Image similarity judging method, electronic equipment and storage medium |
| CN110969165B (en) * | 2019-11-28 | 2024-04-09 | 中国科学院半导体研究所 | Handwritten character recognition method, device, electronic device and storage medium |
| CN111953647B (en) * | 2020-06-22 | 2022-09-27 | 北京百度网讯科技有限公司 | Security verification method and device, electronic equipment and storage medium |
| CN113536771B (en) * | 2021-09-17 | 2021-12-24 | 深圳前海环融联易信息科技服务有限公司 | Element information extraction method, device, equipment and medium based on text recognition |
| CN114067324A (en) * | 2021-10-27 | 2022-02-18 | 厦门友微科技有限公司 | Identifying method, device, equipment and storage medium for identifying verification code picture |
| CN118015634B (en) * | 2024-01-24 | 2024-08-16 | 北京大学 | A data processing system for obtaining verification results of handwritten characters |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104517094A (en) * | 2013-09-30 | 2015-04-15 | 阿里巴巴集团控股有限公司 | Identity authentication method and identity authentication device |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SG121850A1 (en) * | 1999-06-30 | 2006-05-26 | Silverbrook Res Pty Ltd | Method and system for conferencing using sensor with identifier |
| AUPR824601A0 (en) * | 2001-10-15 | 2001-11-08 | Silverbrook Research Pty. Ltd. | Methods and system (npw004) |
| CN103678976A (en) * | 2012-09-18 | 2014-03-26 | 阿里巴巴集团控股有限公司 | Method for authenticating identity through handwriting input device and handwriting input device |
| CN103150019B (en) * | 2013-03-12 | 2016-12-28 | 深圳市国华识别科技开发有限公司 | A kind of hand-written input system and method |
| CN104281831B (en) * | 2013-07-02 | 2017-10-13 | 阿里巴巴集团控股有限公司 | A kind of method and apparatus of person's handwriting checking |
| EP3061067B1 (en) * | 2013-10-25 | 2023-11-29 | Wacom Co., Ltd. | Dynamic handwriting verification, handwriting-baseduser authentication, handwriting data generation, and handwriting data preservation |
| CN104917612B (en) * | 2014-03-14 | 2019-02-12 | 腾讯科技（深圳）有限公司 | Auth method and device |
| CN104239770A (en) * | 2014-09-23 | 2014-12-24 | 陈包容 | Method and system for enabling mobile terminal to enter application program under standby state |
| CN105631272B (en) * | 2016-02-02 | 2018-05-11 | 云南大学 | A kind of identity identifying method of multiple security |
-
2017
- 2017-02-28 CN CN201710113449.2A patent/CN106951832B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104517094A (en) * | 2013-09-30 | 2015-04-15 | 阿里巴巴集团控股有限公司 | Identity authentication method and identity authentication device |
Non-Patent Citations (2)
| Title |
|---|
| tensorflow实现softmax回归（softmax regression）;搬砖小工053;《https://blog.csdn.net/sa14023053/article/details/51884894》;20160711;第1-6页 * |
| 验证码识别与反识别技术;grafx;《https://blog.csdn.net/grafx/article/details/18142589》;20140111;1-9 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106951832A (en) | 2017-07-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106951832B (en) | Verification method and device based on handwritten character recognition | |
| WO2021027336A1 (en) | Authentication method and apparatus based on seal and signature, and computer device | |
| CN107967475B (en) | Verification code identification method based on window sliding and convolutional neural network | |
| CN109993201B (en) | Image processing method, device and readable storage medium | |
| CN110503054B (en) | Method and device for processing text images | |
| CN110598686B (en) | Invoice identification method, system, electronic equipment and medium | |
| CN107358148B (en) | Anti-cheating network investigation method and device based on handwriting recognition | |
| CN104751153B (en) | A kind of method and device of identification scene word | |
| CN113361666B (en) | Handwritten character recognition method, system and medium | |
| CN101452532A (en) | Text-independent handwriting identification method and device | |
| WO2022103564A1 (en) | Fraud detection via automated handwriting clustering | |
| CN104239872A (en) | Abnormal Chinese character identification method | |
| CN114881698A (en) | Advertisement compliance auditing method and device, electronic equipment and storage medium | |
| CN116311214B (en) | License plate recognition method and device | |
| CN112200789B (en) | Image recognition method and device, electronic equipment and storage medium | |
| CN113111882A (en) | Card identification method and device, electronic equipment and storage medium | |
| Ubul et al. | Off-line Uyghur signature recognition based on modified grid information features | |
| CN118072336B (en) | Fixed format card and form structured recognition method based on OpenCV | |
| Aravinda et al. | Template matching method for Kannada handwritten recognition based on correlation analysis | |
| CN114386013A (en) | Automatic student status authentication method and device, computer equipment and storage medium | |
| Thokchom et al. | Recognition of Handwritten Character of Manipuri Script. | |
| CN111259894B (en) | Certificate information identification method and device and computer equipment | |
| CN114332878A (en) | Handwriting processing method and device | |
| Sharma et al. | Primitive feature-based optical character recognition of the Devanagari script | |
| KumarNath et al. | Improving various offline techniques used for handwritten character recognition: a review |
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 |