-
Design And Implementation Of Gabor Filter Based Offline YorÙbà Handwritten Recognition System.
CHAPTER THREE -- [Total Page(s) 3]
Page 2 of 3
-
-
-
3.2.2 NORMALIZATION
Normalization methods are the process of removes all types of variations during the writing and obtains standardized data. This project used slant normalization method. Slant Normalization is one of the measurable factors of different handwriting styles is the angle between longest strokes in a word and the vertical direction referred to the word slant. Slant normalization is used to normalize all characters to a standard form with no slant. The slant detection is performed by dividing the images into different vertical windows through removal of horizontal lines with long runs. These windows are then further divided horizontally by removing window portions that contain no writing. The slant is estimated based on the center of gravity of the upper and lower half of each window averaged over all the windows. And also, another approach was used in which projection profiles are computed for a number of angles away from the vertical direction. The angle corresponding to the projection with the greatest positive derivative is used to detect the least amount of overlap between vertical stroke and therefore the dominant slant angle.
3.2.3 COMPRESSION
In order to reduce storage requirement and increase processing speed, pixelwise Thinning was used, which the process is of extracts the shape information of characters. However during the thinning process a compression in data can be obtained.
Gabor filter image enhancement algorithm is a useful step to be employed prior to minutiae extraction that is capable to enhance the clarity of ridge structure which reduces noise but less effective in low quality images due to in accurate estimation of the orientation and ridge frequency parameter. However, Modified Gabor Filter helps in preserving the handwriting image topology and has an advantage of image- independent but fails when image regions are contaminated with heavy noise also accurate computation of ridges and valley is prohibitively difficult. The algorithm is stated below:
STEP1: All the kernel (GABOR) coefficients are
declared as parameters.p0, p2, p3………. p8
STEP2: When Convolution Signal is 1‟b0 then Filter
data input loaded into memory .i.e.
Memory [8] <=Filter_data_in [0],
Memory [7] <=Filter_data_in [1],
Memory [6] <=Filter_data_in [2],
Memory [0] <=Filter_data_in [8].
It takes 9 clock cycles to finish step 2.
STEP3: When convolution signal is 1‟b1 then the
Convolution process takes place between memory and
Coefficients.
It takes 1 complete clock cycle to finish step 3.
STEP4: When write signal is 1‟b1 then the convolution
output is loaded into the register.
It takes 1 complete clock cycle to finish step 4.
STEP 5: When output_unable is 1‟b1 then the data in
the register is sent to Filter_data_out. When the
convolution signal is 1‟b0 again the process goes to
STEP 2.
It takes 1 complete clock cycle to finish step 5.
CHAPTER THREE -- [Total Page(s) 3]
Page 2 of 3
-
-
ABSRACT - [ Total Page(s): 1 ]ABSTRACT COMING SOON , CHECK OTHERS ... Continue reading---
APPENDIX A - [ Total Page(s): 11 ]s=s+n; e.putString("d"+num,s); e.commit(); new AlertDialog.Builder(MainActivity.this) .setMessage(R.string.learn_sample) .setNeutralButton(R.string.ok,null) .show(); dv.resetPath(); Paths.reset(); dv.invalidate(); } } ... Continue reading---
CHAPTER ONE - [ Total Page(s): 2 ]CHAPTER ONEINTRODUCTION1.1 BACKGROUND OF THE STUDYCharacter is the basic building block of any language which is used to develop different language structures. Characters are alphabets and the structures developed are the words, strings, sentences, paragraphs and so on (Le Cun et al., 1990). Character recognition also known as optical character recognition is the recognition of optically processed characters. The purpose of character recognition is to interpret input as a sequence of chara ... Continue reading---
CHAPTER TWO - [ Total Page(s): 7 ]CHAPTER TWO LITERATURE REVIEW2.1 PATTERN RECOGNITIONPattern recognition is nearly synonymous with machine learning. This branch of artificial intelligence focuses on the recognition of patterns and regularities in data. In many cases, these patterns are learned from labeled "training" data (supervised learning), but when no labeled data are available other algorithms can be used to discover previously unknown patterns (unsupervised learning).The terms ... Continue reading---
CHAPTER FOUR - [ Total Page(s): 4 ]CHAPTER FOURRESULT AND DISCUSSIONS4.1 SYSTEM RESULT ANALYSISBased on the definition given in Handwriting recognition system, 50% of the respondents can be classified as Strong accurate writers, 30% as accurate writers, 15% as Non poor writers and 5% as poor writers. This shows that 95% of handwriting image in the project belong to strong accurate and accurate writers.As far as the gender is concerned, 60% of the respondents were male and 40% were female. This indicates that men are more ap ... Continue reading---
CHAPTER FIVE - [ Total Page(s): 1 ]CHAPTER FIVESUMMARY, CONCLUSION AND RECOMMENDATION5.1 SUMMARYThis project is predicated by the need and necessity to examine the performance evaluation of the Yoruba handwriting image enhancement algorithms. In a bid to achieve this, the Gabor Filter algorithm was used in order to enhance handwriting images so as to test the quality and efficiency recognition.Having implemented this, the levels of performance of the handwriting image enhancement algorithms (Gabor Filter) by comparing the a ... Continue reading---
REFRENCES - [ Total Page(s): 1 ]REFERENCEHuang, B.; Zhang, Y. and Kechadi, M.; Preprocessing Techniques for Online Handwriting Recognition. Intelligent Text Categorization and Clustering, Vol. 164, 2009.J.Pradeep, E.Srinivasanand S.Himavathi, Diagonal based feature extraction for handwritten alphabets recognition System using neural network, Vol 3, No 1, Feb 2011.Jin Chen, Huaigu Cao, Rohit Prasad, Anurag Bhardwaj and Prem Natarajan,Gabor Features for Offline Arabic Handwriting Recognition, 10, June 9-11, 2010.Jumoke F. A ... Continue reading---
