CN105760891A - Chinese character verification code recognition method - Google Patents
Chinese character verification code recognition method Download PDFInfo
- Publication number
- CN105760891A CN105760891A CN201610117882.9A CN201610117882A CN105760891A CN 105760891 A CN105760891 A CN 105760891A CN 201610117882 A CN201610117882 A CN 201610117882A CN 105760891 A CN105760891 A CN 105760891A
- Authority
- CN
- China
- Prior art keywords
- image
- chinese character
- identified
- character
- identifying code
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/63—Scene text, e.g. street names
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Character Discrimination (AREA)
Abstract
The invention relates to a Chinese character verification code recognition method, and the method comprises the following steps: 1), employing a K-means algorithm to carry out the clustering of foreground pixel coordinates, obtaining the coordinates of a mean value point of a two-dimensional Gaussian model, and obtaining the position of a Chinese character in a to-be-recognized image; 2), cutting the to-be-recognized image after the position of the Chinese character is obtained, obtaining the region of the Chinese character in the to-be-recognized image, and extracting the image features of the Chinese character in the region through employing a multi-scale Gabor filtering core; 3), extracting the features of a training set image with a marked character label according to the methods at step 1) and step 2); 4), carrying out the training of a polynomial logistic regression classifier, and predicting the image futures of the Chinese character through employing the polynomial logistic regression classifier, wherein the prediction result is the Chinese character in the to-be-recognized image. Compared with the prior art, the method is accurate in character positioning, is wide in application range, and is high in recognition precision.
Description
Technical field
The present invention relates to a kind of character identifying method, especially relate to the recognition methods of a kind of Chinese character identifying code.
Background technology
2003, the work of Xuewen Wang etc. showed that feature extraction based on Gabor filtering core can effectively solve word
Symbol identification problem.They manually devise Gabor filtering core, in handwritten form for the width of character, principal direction etc. in sample
Chinese character identification aspect achieves extraordinary effect.But their method needs artificial according to sample design filtering core, right
Sample set relies on relatively big, and generalization ability is not strong, and the method proposed can not be transplanted to the character recognition of other pattern well and appoint
In business.
2010, ox is clean utilized the methods such as medium filtering, binaryzation and connected domain analysis to make an uproar Image semantic classification, removal
Point, then by sciagraphy location character, then by the method identification character of template matches, character is limited in 26 English alphabets
And between 10 Arabic numerals.This research finally also developed a system, some can be selected basic by man-machine interactively
Pretreatment, character locating method.
2013, Zheng Xiaofei was primarily upon in identifying code identification, the identification division of single character, have employed " pseudo-two-dimentional hidden
Markov model " image is classified.He points out, template matching method is required for a certain identifying code and sets up relatively standard
ATL, the most just can be higher to such identifying code discrimination, thus adaptability is the best, its anti-deformation nature is relatively in addition
Difference；Charcter topology analytic approach based on architectural feature need not the parameters knowledge being correlated with in advance, if the structure of identifying code
The most fixing；And using the correlation technique of artificial intelligence field, different identifying codes can be reached by the certain sample of re-training
To higher recognition correct rate, there is stronger Stability and adaptability.The most complicated current identifying code character all have adhesion,
Deformation is the most hollow waits the feature stoping segmentation, and its key cracked of this class identifying code is not to identify, and can be accurate
Separating character.Herein for identifying code character tilt, the degree of distortion relatively big, font is not standard letter, so causes
Use traditional template matching method and structured analysis method to be difficult to correctly identify, and hidden Markov model can describe local spy
Levy, overall structure can be represented again, there is certain noise immunity.
2014, Li Kaisheng for the identifying code of different mode select and use matched ambient interferences filter method,
Character segmentation method, character normalization processing method, know the character after normalization by ripe OCR software afterwards
Not.His method needs the identifying code to different mode to apply different methods respectively, and recognition effect is strongly depend on pre-place
Effect after reason and the OCR method used, because if pretreated character picture is still with miscellaneous line, then OCR method
Error result will be given.
Summary of the invention
Defect that the purpose of the present invention is contemplated to overcome above-mentioned prior art to exist and provide a kind of character locating accurately,
The recognition methods of the Chinese character identifying code that applied widely, accuracy of identification is high.
The purpose of the present invention can be achieved through the following technical solutions:
The recognition methods of a kind of Chinese character identifying code, comprises the following steps:
1) cluster by the foreground pixel coordinate in image to be identified, obtain the position of Chinese character in image to be identified
Put；
2) image to be identified after obtaining Chinese character position is carried out cutting, obtain Chinese character in image to be identified
Region, and use Multiscale Gabor Filters core to extract the characteristics of image of Chinese character in region；
3) according to step 1) and 2) method extract the characteristics of image of training set image of reference character label；
4) train multinomial logistic regression grader according to the characteristics of image of training set image, and use multinomial
Logistic returns grader and is predicted the characteristics of image of Chinese character, it was predicted that the result obtained is in image to be identified
Chinese character.
Described step 1) specifically include following steps:
By the coordinate points of foreground pixel points all in image to be identified by the quantity modeling of Chinese character in image to be identified
For multiple dimensional Gaussian mixed models, use maximization likelihood probability K-means algorithm that foreground pixel point coordinates is clustered,
Obtaining the average point coordinates of each dimensional Gaussian mixed model, these average points are the position of Chinese character in image to be identified
Put.
Described step 2) in, described Multiscale Gabor Filters core includes frequency, yardstick and three, direction parameter.
Described step 2) in, the value of described frequency parameter f is 0.05 or 0.25, the value of scale parameter sigma
It is 1 or 3, the value of directioin parameter is 0,OrDescribed Multiscale Gabor Filters core includes 16 filtering
Core.
Described step 2) in characteristics of image include that 16 filtering cores obtain Two-Dimensional Moment after respectively image being carried out convolution
The average of battle array correspondence and variance.
Described step 4) specifically include following steps:
41) characteristics of image of the training set image of the label of reference character extracted is inputted multinomial Logistic to return
Return in grader and be trained；
42) use the multinomial logistic regression grader after training to the figure of Chinese character in image-region to be identified
As feature is predicted；
43) from the highest vector of the middle select probability that predicts the outcome as recognition result.
Compared with prior art, the invention have the advantages that
One, character locating is accurate: clustered foreground point on image by K-means clustering method, thus location character position
Put, be more suitable for solving character locating problem when adjacent character has a small amount of adhesion compared to upright projection method.
Two, applied widely: Gabor filtering core group (Gabor filter banks) " it is exactly to use multiple parameter
Gabor core extracts feature.This method is filtered with the Gabor collecting image of 16 different parameters, finally asks 16 filters
The average of ripple result and variance are as characteristics of image, and compared to the filtering of single Gabor core, the feature of extraction adapts to different big
Live widths little, different, the character picture of different rotary angle.
Three, accuracy of identification is high: use the Logistic of polynomial form to return as grader, with possible on each position
The character occurred is as sample set, after training grader, the character of each position is carried out classification prediction so that classification prediction
The precision that comparison is high can be reached.
Accompanying drawing explanation
Fig. 1 is identifying code image to be identified.
Fig. 2 is pretreated identifying code image.
Fig. 3 is the identifying code image after obtaining location.
Fig. 4 is the image containing 5 characters after cutting.
Detailed description of the invention
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.
Embodiment:
The recognition methods of a kind of Chinese character identifying code, comprises the following steps:
1) coordinate points of foreground pixel points all in image to be identified is built by the quantity of Chinese character in image to be identified
Mould is multiple dimensional Gaussian mixed models, uses maximization likelihood probability K-means algorithm to gather foreground pixel point coordinates
Class, obtains the average point coordinates of each dimensional Gaussian mixed model, and these average points are Chinese character in image to be identified
Position；
2) image to be identified after obtaining Chinese character position is carried out cutting, obtain Chinese character in image to be identified
Region, and use Multiscale Gabor Filters core to extract the characteristics of image of Chinese character, multiple dimensioned Gabor in this method in region
Filtering core (Gabor filter banks) uses 16 Gabor filtering cores, each Gabor filtering core to have three parameters: frequency,
Yardstick (i.e. the yardstick sigma of Gaussian kernel), direction (theta), 16 Gabor filtering cores in this method are by three parameters
Different valued combinations obtain (frequency f=0.05 or 0.25, direction theta=0,OrYardstick sigma
=1 or 3, it is combined into the Gabor filtering core of 2*4*2=16 kind different parameters)；
3) according to step 1) and 2) method extract the characteristics of image of training set image of reference character label；
4) train multinomial logistic regression grader according to the characteristics of image of training set image, and use multinomial
Logistic returns grader and is predicted the characteristics of image of Chinese character, it was predicted that the result obtained is in image to be identified
Chinese character specifically include following steps:
41) characteristics of image of the training set image of the label of reference character extracted is inputted multinomial Logistic to return
Return in grader and be trained；
42) use the multinomial logistic regression grader after training to the figure of Chinese character in image-region to be identified
As feature is predicted；
43) from the highest vector of the middle select probability that predicts the outcome as recognition result.
Identifying code image to be identified is as it is shown in figure 1, as follows to its process identified:
(1) target image is pre-processed, including medium filtering, morphology open and close operator, connected component analysis etc.,
Obtain pretreated image, as shown in Figure 2；
(2) by step 1) pretreated image is carried out character locating, it is verified in code, the position of 5 characters
Coordinate, as shown in Figure 3；
(3) using certain length as character windows radius (such as 13 pixels), with the window of radius length around character position
As character zone, it is syncopated as character, as shown in the red frame in Fig. 3；
(4) after being syncopated as character, obtaining 5 character pictures, each image length and width are the character windows radius into twice
(26 pixel), as shown in Figure 4.It is entered by the input of each character picture to the multinomial logistic regression grader trained
Row prediction, this grader by mark classification and prognostic chart as an equal amount of character picture training obtain.
16 filter result are obtained respectively with image convolution, for each filter result (two dimension with 16 Gabor filtering cores
Matrix), take its average and variance (such as the matrix of 16x16, then these 256 numerical value are averaged and variance).16 filter result
There are 32 numbers, these 32 numerical value are combined as the vector of 32 dimensions, as the feature of this image.
Logistic returns the operation principle of grader: input a series of sample characteristics having marked classification,
The difference that Logistic grader will obtain marking classification with its classification divided according to certain loss assessment criterion, and to damage
Lose the bigger classifier parameters corresponding to feature to punish, this process of continuous iteration, finally make grader classification with
Mark classification is closer to.So obtained Logistic grader can be carried out point for unknown sample feature (not marking classification)
Class.Multinomial logistic regression is that common Logistic returns extensive on multi-class problem, and this method uses multinomial
Logistic return grader input from training set image zooming-out to feature (each training image is extracted by above method
The characteristic vector of 32 dimensions, the image zooming-out of whole training set obtains series of features vector) be trained, then to unknown classification
32 dimensional feature vectors of image zooming-out are predicted, and obtain this image and belong to the probability of each classification (classification as to be judged has
10 classes, then predicting the outcome is 10 dimensional vectors) the highest that of probability is one-dimensional in the vector that predicts the outcome, and corresponding classification is just made
For the grader result to this image prediction.
Claims (6)
1. the recognition methods of a Chinese character identifying code, it is characterised in that comprise the following steps:
1) cluster by the foreground pixel coordinate in image to be identified, obtain the position of Chinese character in image to be identified；
2) image to be identified after obtaining Chinese character position is carried out cutting, obtain the district of Chinese character in image to be identified
Territory, and use Multiscale Gabor Filters core to extract the characteristics of image of Chinese character in region；
3) according to step 1) and 2) method extract the characteristics of image of training set image of reference character label；
4) train multinomial logistic regression grader according to the characteristics of image of training set image, and use multinomial
Logistic returns grader and is predicted the characteristics of image of Chinese character, it was predicted that the result obtained is in image to be identified
Chinese character.
The recognition methods of a kind of Chinese character identifying code the most according to claim 1, it is characterised in that described step 1)
Specifically include following steps:
The coordinate points of foreground pixel points all in image to be identified is modeled as many as the quantity of Chinese character in image to be identified
Individual dimensional Gaussian mixed model, uses maximization likelihood probability K-means algorithm to cluster foreground pixel point coordinates, obtains
The average point coordinates of each dimensional Gaussian mixed model, these average points are the position of Chinese character in image to be identified.
The recognition methods of a kind of Chinese character identifying code the most according to claim 1, it is characterised in that described step 2)
In, described Multiscale Gabor Filters core includes frequency, yardstick and three, direction parameter.
The recognition methods of a kind of Chinese character identifying code the most according to claim 3, it is characterised in that described step 2)
In, the value of described frequency parameter f is 0.05 or 0.25, and the value of scale parameter sigma is 1 or 3, the value of directioin parameter
Be 0,OrDescribed Multiscale Gabor Filters core includes 16 filtering cores.
The recognition methods of a kind of Chinese character identifying code the most according to claim 4, it is characterised in that described step 2)
In characteristics of image include that 16 filtering cores obtain average corresponding to two-dimensional matrix and variance after respectively image being carried out convolution.
The recognition methods of a kind of Chinese character identifying code the most according to claim 1, it is characterised in that described step 4)
Specifically include following steps:
41) characteristics of image of the training set image of the label of reference character extracted is inputted multinomial logistic regression to divide
Class device is trained；
42) use the multinomial logistic regression grader after training special to the image of Chinese character in image-region to be identified
Levy and be predicted；
43) from the highest vector of the middle select probability that predicts the outcome as recognition result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610117882.9A CN105760891A (en) | 2016-03-02 | 2016-03-02 | Chinese character verification code recognition method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610117882.9A CN105760891A (en) | 2016-03-02 | 2016-03-02 | Chinese character verification code recognition method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105760891A true CN105760891A (en) | 2016-07-13 |
Family
ID=56331655
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610117882.9A Pending CN105760891A (en) | 2016-03-02 | 2016-03-02 | Chinese character verification code recognition method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105760891A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106228166A (en) * | 2016-07-27 | 2016-12-14 | 北京交通大学 | The recognition methods of character picture |
| CN106778505A (en) * | 2016-11-24 | 2017-05-31 | 福州瑞芯微电子股份有限公司 | A kind of automated graphics recognize dissemination system and method |
| CN106971150A (en) * | 2017-03-15 | 2017-07-21 | 国网山东省电力公司威海供电公司 | Queuing method for detecting abnormality and device that logic-based is returned |
| CN109697353A (en) * | 2018-11-26 | 2019-04-30 | 武汉极意网络科技有限公司 | A kind of verification method and device for clicking identifying code |
| CN109858432A (en) * | 2019-01-28 | 2019-06-07 | 北京市商汤科技开发有限公司 | Method and device, the computer equipment of text information in a kind of detection image |
| CN110653824A (en) * | 2019-07-26 | 2020-01-07 | 同济人工智能研究院(苏州)有限公司 | Method for characterizing and generalizing discrete trajectory of robot based on probability model |
| CN110837838A (en) * | 2019-11-06 | 2020-02-25 | 创新奇智(重庆)科技有限公司 | End-to-end frame number identification system and method based on deep learning |
| CN110942074A (en) * | 2018-09-25 | 2020-03-31 | 京东数字科技控股有限公司 | Character segmentation recognition method and device, electronic equipment and storage medium |
| CN111259366A (en) * | 2020-01-22 | 2020-06-09 | 支付宝(杭州)信息技术有限公司 | Verification code recognizer training method and device based on self-supervision learning |
| CN112270325A (en) * | 2020-11-09 | 2021-01-26 | 携程旅游网络技术（上海）有限公司 | Character verification code recognition model training method, recognition method, system, device and medium |
| CN112990367A (en) * | 2021-04-25 | 2021-06-18 | 杭州晟视科技有限公司 | Image processing method, device, equipment and storage medium |
| CN113591857A (en) * | 2020-04-30 | 2021-11-02 | 阿里巴巴集团控股有限公司 | Character image processing method and device and ancient Chinese book image identification method |
| CN117011855A (en) * | 2023-10-08 | 2023-11-07 | 深圳市豪斯莱科技有限公司 | Character string image cutting and identifying method, system and readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102043959A (en) * | 2010-12-28 | 2011-05-04 | 青岛海信网络科技股份有限公司 | License plate character segmentation method |
| US20150071528A1 (en) * | 2013-09-11 | 2015-03-12 | Digitalglobe, Inc. | Classification of land based on analysis of remotely-sensed earth images |
| CN104992173A (en) * | 2015-06-03 | 2015-10-21 | 北京好运到信息科技有限公司 | Symbol recognition method and system used for medical report |
-
2016
- 2016-03-02 CN CN201610117882.9A patent/CN105760891A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102043959A (en) * | 2010-12-28 | 2011-05-04 | 青岛海信网络科技股份有限公司 | License plate character segmentation method |
| US20150071528A1 (en) * | 2013-09-11 | 2015-03-12 | Digitalglobe, Inc. | Classification of land based on analysis of remotely-sensed earth images |
| CN104992173A (en) * | 2015-06-03 | 2015-10-21 | 北京好运到信息科技有限公司 | Symbol recognition method and system used for medical report |
Non-Patent Citations (2)
| Title |
|---|
| 唐健炜: ""车牌识别系统的研究"", 《中国优秀硕士学位论文全文数据库 信息科技II辑》 * |
| 张伟: ""汉字笔迹鉴别算法的研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106228166B (en) * | 2016-07-27 | 2019-05-21 | 北京交通大学 | The recognition methods of character picture |
| CN106228166A (en) * | 2016-07-27 | 2016-12-14 | 北京交通大学 | The recognition methods of character picture |
| CN106778505A (en) * | 2016-11-24 | 2017-05-31 | 福州瑞芯微电子股份有限公司 | A kind of automated graphics recognize dissemination system and method |
| CN106778505B (en) * | 2016-11-24 | 2019-09-20 | 福州瑞芯微电子股份有限公司 | A kind of automated graphics identification dissemination system and method |
| CN106971150A (en) * | 2017-03-15 | 2017-07-21 | 国网山东省电力公司威海供电公司 | Queuing method for detecting abnormality and device that logic-based is returned |
| CN110942074A (en) * | 2018-09-25 | 2020-03-31 | 京东数字科技控股有限公司 | Character segmentation recognition method and device, electronic equipment and storage medium |
| CN110942074B (en) * | 2018-09-25 | 2024-04-09 | 京东科技控股股份有限公司 | Character segmentation recognition method and device, electronic equipment and storage medium |
| CN109697353A (en) * | 2018-11-26 | 2019-04-30 | 武汉极意网络科技有限公司 | A kind of verification method and device for clicking identifying code |
| CN109858432B (en) * | 2019-01-28 | 2022-01-04 | 北京市商汤科技开发有限公司 | Method and device for detecting character information in image and computer equipment |
| CN109858432A (en) * | 2019-01-28 | 2019-06-07 | 北京市商汤科技开发有限公司 | Method and device, the computer equipment of text information in a kind of detection image |
| CN110653824A (en) * | 2019-07-26 | 2020-01-07 | 同济人工智能研究院(苏州)有限公司 | Method for characterizing and generalizing discrete trajectory of robot based on probability model |
| CN110837838A (en) * | 2019-11-06 | 2020-02-25 | 创新奇智(重庆)科技有限公司 | End-to-end frame number identification system and method based on deep learning |
| CN111259366A (en) * | 2020-01-22 | 2020-06-09 | 支付宝(杭州)信息技术有限公司 | Verification code recognizer training method and device based on self-supervision learning |
| CN113591857A (en) * | 2020-04-30 | 2021-11-02 | 阿里巴巴集团控股有限公司 | Character image processing method and device and ancient Chinese book image identification method |
| CN113591857B (en) * | 2020-04-30 | 2024-12-13 | 阿里巴巴集团控股有限公司 | Character image processing method, device and ancient Chinese book image recognition method |
| CN112270325A (en) * | 2020-11-09 | 2021-01-26 | 携程旅游网络技术（上海）有限公司 | Character verification code recognition model training method, recognition method, system, device and medium |
| CN112270325B (en) * | 2020-11-09 | 2024-05-24 | 携程旅游网络技术（上海）有限公司 | Character verification code recognition model training method, recognition method, system, equipment and medium |
| CN112990367A (en) * | 2021-04-25 | 2021-06-18 | 杭州晟视科技有限公司 | Image processing method, device, equipment and storage medium |
| CN117011855A (en) * | 2023-10-08 | 2023-11-07 | 深圳市豪斯莱科技有限公司 | Character string image cutting and identifying method, system and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105760891A (en) | Chinese character verification code recognition method | |
| Saxena | Niblack’s binarization method and its modifications to real-time applications: a review | |
| Yi et al. | Localizing text in scene images by boundary clustering, stroke segmentation, and string fragment classification | |
| Lawgali | A survey on Arabic character recognition | |
| CN106384112A (en) | Rapid image text detection method based on multi-channel and multi-dimensional cascade filter | |
| Shaw et al. | Offline Handwritten Devanagari Word Recognition: A holistic approach based on directional chain code feature and HMM | |
| Afakh et al. | Aksara jawa text detection in scene images using convolutional neural network | |
| Roy et al. | Date-field retrieval in scene image and video frames using text enhancement and shape coding | |
| Lawgali et al. | A Frame Work For Arabic Handwritten Recognition Based on Segmentation | |
| Kumar | A study for handwritten Devanagari word recognition | |
| Kim et al. | Text detection with deep neural network system based on overlapped labels and a hierarchical segmentation of feature maps | |
| BJ et al. | Kannada confusing character recognition and classification using random forest and SVM | |
| Feild et al. | Scene text recognition with bilateral regression | |
| Abdo et al. | An approach to analysis of Arabic text documents into text lines, words, and characters | |
| Barnouti et al. | An efficient character recognition technique using K-nearest neighbor classifier | |
| Zhang et al. | OCR with the Deep CNN Model for Ligature Script‐Based Languages like Manchu | |
| Zuo et al. | An intelligent knowledge extraction framework for recognizing identification information from real-world ID card images | |
| KR20200068073A (en) | Improvement of Character Recognition for Parts Book Using Pre-processing of Deep Learning | |
| Ali et al. | Different handwritten character recognition methods: a review | |
| Jumari et al. | A survey and comparative evaluation of selected off-line Arabic handwritten character recognition systems | |
| Chaturvedi et al. | Automatic license plate recognition system using surf features and rbf neural network | |
| Boussellaa et al. | Unsupervised block covering analysis for text-line segmentation of Arabic ancient handwritten document images | |
| Ali et al. | A new design based-fusion of features to recognize Arabic handwritten characters | |
| Kumar et al. | Scene text recognition using artificial neural network: a survey | |
| Lamba et al. | Handwriting Recognition System-A Review |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160713 |
|
| RJ01 | Rejection of invention patent application after publication |