Volume & Issue no: Volume 6, Issue 1, January - February 2017
____________________________________________________________________________________________________
Title: |
Facial Semantics Recognition Method for Content based Video Retrieval Systems |
Author Name: |
B. S. Daga, A. A. Ghatol ,V.M.Thakare |
Abstract: |
Abstract
With growing video databases, the accurate facial expression
identifier systems are proving their importance. State of the
art literature suggests the use of facial texture pattern while
identifying the facial expression from a image frame or
photograph, and exercise with moving geometric patterns
while dealing with dynamics expression identification from a
video. Accordingly they make use of local binary patterns
(LBP) from a image frame or facial landmark tacking (FLT)
to extract out the semantics from a video database. Actual
classification and identification of expression is then
performed by support vector machine (SVM) based classifier.
This work primarily presents a comparative assessment of
LBP and FLT methods as semantic means for video retrieval
systems. Moreover herein introduces possible implication of
probabilistic neural network (PNN) as more effective
alternative to conventionally practiced SVM for classification
problem. The modular system so developed has been tested
with well-established database and comparative results of the
methods are presented. Results presented indicate that the
implication of order of magnitude faster PNN can be efficient
replacement of SVM. Moreover, considering the near future
technologies, use of facial landmark tracking technique is
most viable solution to yield accurate and meaningful results.
The facial expression identification based on different facial
semantics has been one of well-researched areas for quite
some time. However, expecting a proven system implication
into real practice needs to focus on systems performance and
processing speed. The usefulness of current work lies in its
contribution towards presenting a comparative study of
classifiers for content based video retrieval systems.
Keywords: Facial Landmark Tracking, Support Vector
Machine, Neural Network, Semantic, Facial Expressions,
Content Based Video Retrieval System |
Cite this article: |
B. S. Daga, A. A. Ghatol ,V.M.Thakare , "
Facial Semantics Recognition Method for Content based Video Retrieval Systems" , International Journal of Emerging Trends & Technology in Computer Science (IJETTCS) ,
Volume 6, Issue 1, January - February 2017 , pp.
139-146 , ISSN 2278-6856.
|
| Full Text [PDF] Back to Current Issue | |
NOTE: Authors note that paper cannot be withdrawn at any condition once it is accepted. The Team of IJETTCS advise you, do not submit same article to the multiple journals simultaneously. This may create a problem for you. Please wait for review report which will take maximum 01 to 02 week.