Color Textured Feature based Image retrieval using Local Binary Pattern with Hyper plane Thresholding
C.CallinsChristiyana1, M. Mathinakani2, M. Poomani Punitha3, K. Priyadharsini4
1C. Callins Christiyana, Professor, Computer Science and Engineering, Sethu Institute of Technology.
2M .Mathinakani, Assistant Professor (S.G), Computer Science and Engineering, Sethu Institute of Technology.
3M. Poomani Punitha, Asst. Professor (S.G), Computer Science and Engineering, Sethu Institute of Technology.
4K. Priya dharsini, Asst. Professor (S.G), Computer Science and Engineering, Sethu Institute of Technology.
Manuscript received on 02 June 2019 | Revised Manuscript received on 10 June 2019 | Manuscript published on 30 June 2019 | PP: 912-918 | Volume-8 Issue-8, June 2019 | Retrieval Number: G5517058719/19©BEIESP
Open Access | Ethics and 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: The color image retrieval is a dynamic research area for more than a few decades as to access the images from the outsized image database. The color feature extraction and representation is the vital task in color image retrieval. Among the various color feature representation methods, the retrieval performance of Local Binary Pattern meant for color images (LBPC) is more as it combines color and texture features. LBPC utilizes the definition of hyper_plane to threshold the color pixel as either ‘0’ or ‘1’. The definition of hyper_plane is derived using different normal vectors. This work considers this factor and design LBPC based image retrieval system with three hyper_plane normal vectors; Local average Normal, Center Normal, Mean Normal. Experimental results show that the LBPC with Local average Normal vector hyper_plane and LBPC with Mean Normal vector hyper_plane are yielding the same retrieval efficiency. The performance of Center Normal vector hyper_planed LBPC is low as compared to other two. This work proposed that the LBPC can be extracted using either Local average Normal vector hyper_plane or Mean Normal vector hyper_plane for the efficient retrieval of color images.
Keyword: Content_Based Image Retrieval, LBP, color image, Hyper_plane normal vector, Precision, Recall, MAP,Wang Database.
Scope of the Article: Image analysis and Processing.