Face Recognition Method using Mean-Shift by Means of Region Merging
Najmus Sehar1, Santosh Kushwaha2, Yogesh Rai3
1Najmus Sehar, Shree Institute of Science & Technology, Bhopal (Madhya Pradesh). India.
2Santosh Kushwaha, Shree Institute of Science & Technology, Bhopal (Madhya Pradesh). India.
3Yogesh Rai, Shree Institute of Science & Technology, Bhopal (Madhya Pradesh). India.
Manuscript received on 15 April 2016 | Revised Manuscript received on 25 April 2016 | Manuscript Published on 30 April 2016 | PP: 112-118 | Volume-5 Issue-4, April 2016 | Retrieval Number: D4514045416/16©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: We projected a original method for face matching from face image database. In this technique we have used set of face images because recognition conclusion is based on comparisons of face image database. In this paper we have presented an approach to region based face matching. Here the mean shift low level image segmentation method is used to segment the image into many small regions. As a well-liked segmentation scheme for color image, watershed has over segmentation as compared to mean-shift and also mean-shift conserves edge information of the object very well. The proposed technique mechanically merges the regions that are initially segmented by mean shift segmentation, effectively takes out the object contour and then, matches the obtained mask with test database image sets on the basis of color and texture. Extensive research are performed and the outcome shows that the projected method can reliably figure out the mask from the face image and efficiently matches the mask with face image sets.
Keywords: Face Matching, Image Segmentation, Region Merging, Watershed, Mean shift
Scope of the Article: Image Processing and Pattern Recognition