Research Project
UGC Major Project No. F 41-625/2012 (SR)
Role: Co-Investigator
Period: July 2012 to July 2015 ( 3 Years)
Project Title: Face recognition based driver's alert system
Department: Department of Computer Science & Engineering
University: Manonmaniam Sundaranar University, Tirunelveli
Academic Projects
Ø A Performance Study on Improved Edge Detection approaches for Object Extractions
The objective of this project is to apply the different edge detection methods like Canny, Canny with scale multiplied and Ant colony optimization is applied and their detection level is identified and result shows that canny with scale multiplied provides better result than others and how this technique improves the edge detection rate is studied.
Ø Image segmentation by maximal similarity based region merging and dynamic growth implementation for natural images
In the segmentation approach initial segmentation is done using mean shift, watershed and quadtree segmentation methods. After the initial segmentation process next process is to merge the image based on the user input given, here markers are used to identify the region for object and background. The process follows the interactive scheme for providing input markers on the initial segmented image. Based on the marking the object and background are identified and the relative regions are merged back based on distance measures. Here three different distance measures are used for similarity measures. Bhattacharyya distance, Euclidean Distance and Mahalanobis distance are used for distance measure. The objective of this project is to apply the initial segmentation techniques like mean shift, watershed and quadtree prior to merging process. For merging the regions different similarity measures are used. An interactive user input scheme for segmentation is developed and accuracy is studied on different methods.
Sponsored Projects
(Under Multimedia Research Lab funded by AICTE AQIS MODROBS 2018 )
Objective: The students and staff will be equipped with coding and express their ideas as animated & 3D model.
Current Projects :
(Multimedia perspective)
All the 3D model related to engineering disciples are created using Autodesk 3ds Max and Maya at the Multimedia Research Lab.
(Research perspective)
Blood cell 3D segmentation and 3D profiling done using ImageJ and also for analysing of images in batch processing. For finding deep learning-based human emotion detection and analysis to understand people's response through twitter with the packages like nltk, pandas, re, scikit-learn, numpy, matplotlib in anaconda python in ubuntu distribution .