Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/310951
Title: Pattern Recognition in Human Behavior Analysis Image Analysis Using Artificial Neural Network or Any Other Methods
Researcher: Syeda Asra
Guide(s): Shubhangi, D. C
Keywords: Computer Science
Computer Science Artificial Intelligence
Engineering and Technology
University: Visvesvaraya Technological University, Belagavi
Completed Date: 2018
Abstract: The human behavior recognition is a significant task of discovery and identification of newlinebehaviors through an input hand written image is challenging. The essence of human newlinebehaviour recognition is to find what the writer is all about and is not the writer s newlineidentification i-e the analysis answers the question what is he, rather than who is he. A newlinesignature is scanty text which is written on a paper and comes from a conscious mind. The newlinetrue personality profile cannot be analyzed by taking signature samples for analysis. newlineThe human behaviour is influenced by the inherent personality trait which the human newlinepossess. This work is a fusion of psychology and image processing. The problem was newlineidentified in the domain of psychology and implement by using image processing techniques. newlineA paragraph comprising of three was given to write for the individuals. This was done with newlinean intent that the sub conscious mind comes into play after writing first paragraph. Data set newlineconsisted 500 samples equally of male and female. In order to make the system robust and newlinetime, mood, gender and age invariant, the samples were collected at different point of time, newlinedays and people belonging to the age between 18 to 80.Various feature extraction techniques newlinelike histogram orientation gradients, Covo-Co-HOG, rotation, scale and translation invariant newlinetechniques, edge histogram descriptor, Zernike moments and hierarchical centroid based newlineshape descriptor were employed. newlineSegmentation methods like drop fall algorithm, region of interest, free chain coding newlinewere employed. Letters like O, T, I and direction of handwriting samples like uphill downhill newlineand constant were considered in order to find the pessimism; optimism and self-controlled newlinebehaviour in an individual, further left margins, right margins etc. were used to analyze newlinepersons inclination in past, future also person being economical, is revealed. Two phases newlineknown as testing and training were employed. The feature extracted and trained using SVM newlineclassifier.
Pagination: VII, 101
URI: http://hdl.handle.net/10603/310951
Appears in Departments:Department of Computer Science and Engineering



Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.

Altmetric Badge: