AU2021103831A4 - Method and system for Font Character Recognition - Google Patents
Method and system for Font Character Recognition Download PDFInfo
- Publication number
- AU2021103831A4 AU2021103831A4 AU2021103831A AU2021103831A AU2021103831A4 AU 2021103831 A4 AU2021103831 A4 AU 2021103831A4 AU 2021103831 A AU2021103831 A AU 2021103831A AU 2021103831 A AU2021103831 A AU 2021103831A AU 2021103831 A4 AU2021103831 A4 AU 2021103831A4
- Authority
- AU
- Australia
- Prior art keywords
- character image
- character
- recognition
- image
- feature
- 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.)
- Ceased
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000000605 extraction Methods 0.000 claims abstract description 21
- 230000008569 process Effects 0.000 claims description 12
- 238000013527 convolutional neural network Methods 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 4
- 238000007637 random forest analysis Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 230000000306 recurrent effect Effects 0.000 claims description 3
- 238000012706 support-vector machine Methods 0.000 claims description 3
- 230000008901 benefit Effects 0.000 description 9
- 238000010586 diagram Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 2
- 238000012015 optical character recognition Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/243—Aligning, centring, orientation detection or correction of the image by compensating for image skew or non-uniform image deformations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06V30/22—Character recognition characterised by the type of writing
- G06V30/226—Character recognition characterised by the type of writing of cursive writing
- G06V30/2268—Character recognition characterised by the type of writing of cursive writing using stroke segmentation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Character Discrimination (AREA)
Abstract
The present invention generally relates to Method and system for Font
Character Recognition. The font recognition method comprises, acquiring a
text image comprising one or more words each including one or more
characters, acquiring a character image by segmenting the text image into
separate images of the characters, digitizing the character image to form a
two-dimensional array of pixels, performing feature extraction by generating
a feature vector corresponding to the character image, wherein the feature
vector is a set of values corresponding to attributes and characteristics of a
character and feeding the feature vector into a recognition network to
recognize and determine a class label for the character image, wherein the
class label is one of a predetermined set of uppercase letters, lowercase
letters and digits.
15
LM
c- E 4 42 o w
E aE
- 0
40- E 0
4e4
00
Eu
E0 0
E 4
L 4
4
Description
E 42 o c- 4 w
E aE - 0
40- E 0 4e4
00
Eu
E0 0
E 4
4 L 4
Method and system for Font Character Recognition
The present invention relates to a character recognition system. In particular, the present invention relates toMethod and system for Font Character Recognition.
Optical Character Recognition is used in identification of a license plate, recognition from documents, and recognition of street numbers. Offline OCR technologies rely on the text identification of character as well as word from document images and natural scenes. Recognizingcomputer-synthesized font characters in offline is essential as characterscan be in different fonts as well as styles.Such forms of characters are often very hard to identify, and sometimes it takes along time to translate them into machine-understandable text forms. In the view of the forgoing discussion, it is clearly portrayed that there is a need to have a system and method for font character recognition system with improved feature extraction. The proposed invention provides the best accuracy when compared with existing recognition systems. It uses an improved feature extraction algorithm to achieve this accuracy.
The present disclosure seeks to provide a system and a method for font character recognition system and method with improved feature extraction.This system can recognize text from a document image. The system operates in 3 stages - preprocessing, extraction of features, and recognition or classification. In the phase of preprocessing - binarization, thinning and edge detection techniques are used. In the process of recognition, we are assigning correct classlabels to the test sample based on the features derived from the training data in the feature extraction step.
In an embodiment a font character recognition system with improved feature extraction, comprises of at least one processor. The system further comprises a memory including instructions that, when executed by the at least one processor, cause the system toacquire a text image comprising one or more words each including one or more characters, acquire a character image by segmenting the text image into separate images of the characters, digitize the character image to form a two-dimensional array of pixels each associated with a pixel value, wherein the pixel value is expressed in a binary number, pre-process the character image, extract features by generating a feature vector corresponding to the character image, wherein the feature vector is a set of values corresponding to attributes and characteristics of a character andfeed the feature vector into a recognition network to recognize and determine a class label for the character image, wherein the class label is one of a predetermined set of uppercase letters, lowercase letters and digits.
In another embodiment, the feature extractor of the system is configured through one of the following methods: angular motion of shape based feature (AMSF) method, distance from center to thinned lined features (DCTF) method, distance from center to outer edge features (DCEF) method, or a combination thereof.
In another embodiment, the recognition network can be a support vector machine, a random forest, a recurrent neural network, a convolutional neural network, or a combination thereof.
In another embodiment, the recognition network is a convolutional neural network configured by 2 dense layers of convolution, a rectified linear unit(ReLU) function and a SoftMax activation function.
To further clarify advantages and features of the present disclosure, a more particular description of the invention will be rendered by reference to specific embodiments thereof, which is illustrated in the appended drawings. It is appreciated that these drawings depict only typical embodiments of the invention and are therefore not to be considered limiting of its scope. The invention will be described and explained with additional specificity and detail with the accompanying drawings.
These and other features, aspects, and advantages of the present disclosure will become better understood when the following detailed description is read with reference to the accompanying drawings in which like characters represent like parts throughout the drawings, wherein:
Figure 1 illustrates a block diagram of font character recognition system with improved feature extraction in accordance with an embodiment of the present disclosure.
Figure 2 illustrates a flow chart of a font character recognition method with improved feature extraction in accordance with an embodiment of the present disclosure.
Figure 3 illustrates a flow chart of aprocess of offline character recognitionin accordance with an embodiment of the present disclosure.
Figure 4 illustrates a block diagram of font character recognition system with improved feature extraction in accordance with an embodiment of the present disclosure.
Figure 5 illustrates the process of unique encrypted code generation by a sliding window in accordance with an embodiment of the present disclosure.
Figure 6 illustrates the process of AMSF type feature extraction using a sliding window in accordance with an embodiment of the present disclosure.
Figure 7 illustrates a table depicting the recognition accuracy percentage of the present system when used on Chars74kFnt whole character dataset.
Figure 8 illustrates a graph depicting the recognition accuracy percentage of the present system when CNN classifier and Chars74kFnt whole character dataset is used.
Figure 9 illustrates a table depicting a comparison among the present system and existing font recognition systems.
Further, skilled artisans will appreciate that elements in the drawings are illustrated for simplicity and may not have necessarily been drawn to scale. For example, the flow charts illustrate the method in terms of the most prominent steps involved to help to improve understanding of aspects of the present disclosure. Furthermore, in terms of the construction of the device, one or more components of the device may have been represented in the drawings by conventional symbols, and the drawings may show only those specific details that are pertinent to understanding the embodiments of the present disclosure so as not to obscure the drawings with details that will be readily apparent to those of ordinary skill in the art having benefit of the description herein.
For the purpose of promoting an understanding of the principles of the invention, reference will now be made to the embodiment illustrated in the drawings and specific language will be used to describe the same. It will nevertheless be understood that no limitation of the scope of the invention is thereby intended, such alterations and further modifications in the illustrated system, and such further applications of the principles of the invention as illustrated therein being contemplated as would normally occur to one skilled in the art to which the invention relates.
It will be understood by those skilled in the art that the foregoing general description and the following detailed description are exemplary and explanatory of the invention and are not intended to be restrictive thereof.
Reference throughout this specification to "an aspect", "another aspect" or similar language means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. Thus, appearances of the phrase "in an embodiment", "in another embodiment" and similar language throughout this specification may, but do not necessarily, all refer to the same embodiment.
The terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process or method that comprises a list of steps does not include only those steps but may include other steps not expressly listed or inherent to such process or method. Similarly, one or more devices or sub-systems or elements or structures or components proceeded by "comprises...a" does not, without more constraints, preclude the existence of other devices or other sub systems or other elements or other structures or other components or additional devices or additional sub-systems or additional elements or additional structures or additional components.
Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The system, methods, and examples provided herein are illustrative only and not intended to be limiting.
Embodiments of the present disclosure will be described below in detail with reference to the accompanying drawings.
Figure 1 illustrates a block diagram of font character recognition system with improved feature extraction in accordance with an embodiment of the present disclosure. A font character recognition system 100 with improved feature extractioncomprises ofat least one processor 102. The system 100 further comprises a memory 104 including instructions that, when executed by the at least one processor, cause the system to acquire a text image comprising one or more words each including one or more characters.The system 100 then acquires a character image by segmenting the text image through a segmentor 106 into separate images of the characters. The system 100 then digitizes the character image to form a two-dimensional array of pixels each associated with a pixel value, wherein the pixel value is expressed in a binary number.The system then pre-processes the character image through a pre-processor 108. It then extracts features through a feature extractor 110 by generating a feature vector corresponding to the character image, wherein the feature vector is a set of values corresponding to attributes and characteristics of a character. Lastly the system 100feeds the feature vector into a recognition network 112 to recognize and determine a class label for the character image, wherein the class label is one of a predetermined set of uppercase letters, lowercase letters and digits.
Figure 2 illustrates a flow chart of a font character recognition method with improved feature extraction in accordance with an embodiment of the present disclosure
At step 202, the font recognition method 200 comprisesacquiring a text image comprising one or more words each including one or more characters. At step 204, the font recognition method 200 comprisesacquiring a character image by segmenting the text image into separate images of the characters. At step 206, the font recognition method 200 comprises digitizing the character image to form a two-dimensional array of pixels each associated with a pixel value, wherein the pixel value is expressed in a binary number. At step 208, the font recognition method 200 comprisespre-processing the character image. At step 210, the font recognition method 200 comprises performing feature extraction by generating a feature vector corresponding to the character image, wherein the feature vector is a set of values corresponding to attributes and characteristics of a character. At step 212, the font recognition method 200 comprisesfeeding the feature vector into a recognition network to recognize and determine a class label for the character image, wherein the class label is one of a predetermined set of uppercase letters, lowercase letters and digits.
In another embodiment, the feature extractor of the system is configured through one of the following methods: angular motion of shape based feature (AMSF) method, distance from center to thinned lined features (DCTF) method, distance from center to outer edge features (DCEF) method, or a combination thereof.
In another embodiment, the recognition network can be a support vector machine, a random forest, a recurrent neural network, a convolutional neural network, or a combination thereof.
In another embodiment, the recognition network is a convolutional neural network configured by 2 dense layers of convolution, a rectified linear unit(ReLU) function and a SoftMax activation function.
Figure 3 illustrates a flow chart of a process of offline character recognition in accordance with an embodiment of the present disclosure.
Figure 4 illustrates a block diagram of font character recognition system with improved feature extraction in accordance with an embodiment of the present disclosure.
Figure 5 illustrates the process of unique encrypted code generation by a sliding window in accordance with an embodiment of the present disclosure.
Figure 6 illustrates the process of AMSF type feature extraction using a sliding window in accordance with an embodiment of the present disclosure.
The system was implemented onChars74K dataset. This dataset has 74 thousand character images. The dataset contains 62 distinctclasses. Those are digits, uppercase alphabets, and lowercase alphabets. The dataset having 7705 natural character images, 3401 hand-drawn characters using PC and 62,992 computer-synthesized fonts. The Font sub-dataset (chars74kFnt) consists of black and white characters of 128 x 128 resolution, with 4 variations. Those are of the italic, bold, regular and bold with italic combinations. Every character has 1016 versions. The recognition is carriedout on a 75:25 ratio of data to compose training and testing data. The total number of samples is 62,992. Of which 10,160 digit samples and 26,416 samples each for uppercase and lowercase.
The system was implemented by three different extraction procedures of the features- AMSF, DCTF, DCEF. The AMSF algorithm positions the current text pixel in the center of the sliding window and checks its neighboring location. This encryption dependsabsolutely on the pointed position of the neighboring pixels in the sliding window. There are 9 cells on the sliding window. Since the current pixel occupies acentered cell and then the other 8 cells help the feature vector to be created. We have 256, which are 28 number of possibilities of different distinctive feature values. The codes are created only in one direction, from left end to the right end point of the sliding window. Thus, the optimal combination number of symbols is 256/2 = 128. The sliding window scans the image into two zones, one in the lower zone and the other in the upper zone. Therefore, the total combination of encrypted symbol is 128 + 128 = 256. The DCTF algorithm takes in considerations the shape of different characters. The distance from the center of these different text images is different. Based on this definition, the distance is computed from the central.
The distance from the middle location to its thin text outline is determined in 8 directions. Intersection points are the positions on a binarized image, where each of eight directions finds the first white pixel in their respective paths. This length is the Euclidean's distance of the angular position and is a floating value. The DCTF feature matrix stores these values sequentially. Distance from Center to the Outer Edge Feature (DCEF) method is based on the distance from the outer edge to the middle of the character image. Every eight angular distances from the center are computed according to the cell positions to generate the DCEF feature matrix. Thereare two thin lines for an edged version of the image sample. So, always the outer edge will be encountered first. The distance will be from the outer edge to its center.
For classification and recognition, four machine learning approaches are implemented. Those are SVM Random forest, RNN, and CNN. The results were captured for various combinations of feature extraction methods and classification networks, based on average of multiple runs. It was observed that the font character recognition system performs better in the case of digit recognition. CNN classifier showed better visibility of the recognition rate on train data and test data to the number of the epoch. The recognition accuracy of the system in the case of upper and lower case alphabets are also fair enough when compared to other existing font recognition systems.The best accuracy of the system is achieved when the combination of three feature extraction methods is used with deep-learning CNN classifiers.
Figure 7 illustrates a table depicting the recognition accuracy percentage of the present system when used on Chars74kFnt whole character dataset.
Figure 8 illustrates a graph depicting the recognition accuracy percentage of the present system when CNN classifier and Chars74kFnt whole character dataset is used.
Figure 9 illustrates a table depicting a comparison among the present system and existing font recognition systems.
The drawings and the forgoing description give examples of embodiments. Those skilled in the art will appreciate that one or more of the described elementsmay well be combined into a single functional element. Alternatively, certain elementsmay be split into multiple functional elements. Elementsfrom one embodiment may be added to another embodiment. For example, orders of processes described herein may be changed and are not limited to the manner described herein. Moreover, the actions of any flow diagram need not be implemented in the order shown; nor do all of the acts necessarily need to be performed. Also, those acts that are not dependent on other acts may be performed in parallel with the other acts. The scope of embodiments is by no meanslimited by these specific examples. Numerous variations, whether explicitly given in the specification or not, such as differences in structure, dimension, and use of material, are possible. The scope of embodiments is at least as broad as given by the following claims.
Benefits, other advantages, and solutions to problems have been described above with regard to specific embodiments. However, the benefits, advantages, solutions to problems, and any component(s) that may cause any benefit, advantage, or solution to occur or become more pronounced are not to be construed as a critical, required, or essential feature or component of any or all the claims.
Claims (7)
1. A font character recognition system with improved feature extraction, comprising: at least one processor; and a memory including instructions that, when executed by the at least one processor, cause the system to: acquire a text image comprising one or more words each including one or more characters; acquire a character image by segmenting the text image into separate images of the characters; digitize the character image to form a two-dimensional array of pixels each associated with a pixel value, wherein the pixel value is expressed in a binary number; pre-process the character image; extract features by generating a feature vector corresponding to the character image, wherein the feature vector is a set of values corresponding to attributes and characteristics of a character; and feed the feature vector into a recognition network to recognize and determine a class label for the character image, wherein the class label is one of a predetermined set of uppercase letters, lowercase letters and digits.
2. The system as claimed in claim 1, wherein the pre-process step executed by at least one processor comprises of: a. obtaining a binary version of the character image using a global thresholding technique; b. thinning of the character image; and c. detecting edges of character in the character image by calculating a Gaussian gradient, wherein steps a-c can occur in a random manner.
3. The system as claimed in claim 1, wherein the extract features step executed by at least one processor is configured through one of the following methods: angular motion of shape based feature(AMSF) method, distance from center to thinned lined features(DCTF) method, distance from center to outer edge features (DCEF) method, or a combination thereof.
4. The system as claimed in claim 1, wherein the recognition network being a support vector machine, a random forest, a recurrent neural network, a convolutional neural network, or a combination thereof.
5. The system as claimed in claim 1 and 2, wherein the recognition network is a convolutional neural network configured by 2 dense layers of convolution, a rectified linear unit(ReLU) function and a SoftMax activation function.
6. A method for font character recognition with improved feature extraction, comprising: a. acquiring a text image comprising one or more words each including one or more characters; b. acquiring a character image by segmenting the text image into separate images of the characters; c. digitizing the character image to form a two-dimensional array of pixels each associated with a pixel value, wherein the pixel value is expressed in a binary number; d. pre-processingthe character image; e. performing feature extractionby generating afeature vector corresponding to the character image, wherein the feature vector is a set of values corresponding to attributes and characteristics of a character; and f. feeding the feature vector into a recognition network to recognize and determine a class label for the character image, wherein the class label is one of a predetermined set of uppercase letters, lowercase letters and digits, wherein the steps a-f occur in a sequential manner
7. The method as claimed in claim 2, wherein feeding the feature vectorinto a recognition network, results in determining that the character image is a specific characterfrom a predetermined set of uppercase letters, lowercase letters and digits.
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| AU2021103831A AU2021103831A4 (en) | 2021-07-02 | 2021-07-02 | Method and system for Font Character Recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| AU2021103831A AU2021103831A4 (en) | 2021-07-02 | 2021-07-02 | Method and system for Font Character Recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| AU2021103831A4 true AU2021103831A4 (en) | 2021-08-26 |
Family
ID=77369681
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| AU2021103831A Ceased AU2021103831A4 (en) | 2021-07-02 | 2021-07-02 | Method and system for Font Character Recognition |
Country Status (1)
| Country | Link |
|---|---|
| AU (1) | AU2021103831A4 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114357204A (en) * | 2021-11-25 | 2022-04-15 | 腾讯科技（深圳）有限公司 | Media information processing method and related equipment |
-
2021
- 2021-07-02 AU AU2021103831A patent/AU2021103831A4/en not_active Ceased
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114357204A (en) * | 2021-11-25 | 2022-04-15 | 腾讯科技（深圳）有限公司 | Media information processing method and related equipment |
| CN114357204B (en) * | 2021-11-25 | 2024-03-26 | 腾讯科技（深圳）有限公司 | Media information processing method and related equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sahare et al. | Multilingual character segmentation and recognition schemes for Indian document images | |
| Kumar et al. | A detailed review of feature extraction in image processing systems | |
| Dave | Segmentation methods for hand written character recognition | |
| Farooq et al. | Pre-processing methods for handwritten Arabic documents | |
| Namboodiri et al. | Document structure and layout analysis | |
| Zayene et al. | Multi‐dimensional long short‐term memory networks for artificial Arabic text recognition in news video | |
| Mahto et al. | Combined horizontal and vertical projection feature extraction technique for Gurmukhi handwritten character recognition | |
| Chandio et al. | Character classification and recognition for Urdu texts in natural scene images | |
| Zoizou et al. | A new hybrid method for Arabic multi-font text segmentation, and a reference corpus construction | |
| Amrouch et al. | Handwritten amazigh character recognition system based on continuous hmms and directional features | |
| Tavoli et al. | Statistical geometric components of straight lines (SGCSL) feature extraction method for offline Arabic/Persian handwritten words recognition | |
| AU2021103831A4 (en) | Method and system for Font Character Recognition | |
| Alghyaline | Arabic Optical Character Recognition: A Review. | |
| US10217020B1 (en) | Method and system for identifying multiple strings in an image based upon positions of model strings relative to one another | |
| Shahin | Printed Arabic text recognition using linear and nonlinear regression | |
| Gattal et al. | Segmentation and recognition strategy of handwritten connected digits based on the oriented sliding window | |
| Omidyeganeh et al. | A new segmentation technique for multi font farsi/arabic texts | |
| Ali et al. | UOCR: A ligature based approach for an Urdu OCR system | |
| JP6310155B2 (en) | Character recognition device, character recognition method, and character recognition program | |
| Shanjana et al. | Character segmentation in Malayalam Handwritten documents | |
| Omachi et al. | Structure extraction from decorated characters using multiscale images | |
| Zahid Hasan et al. | A CNN-based classification model for recognizing visual bengali font | |
| Dinges et al. | Automatic recognition of common Arabic handwritten words based on OCR and N-GRAMS | |
| Adhvaryu | Optical character recognition using template matching (Alphabet & Numbers) | |
| Kaur et al. | Adverse conditions and techniques for cross-lingual text recognition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FGI | Letters patent sealed or granted (innovation patent) | ||
| MK22 | Patent ceased section 143a(d), or expired - non payment of renewal fee or expiry |