Classification of Color Textures Using Region Based Motif and Color Features
K.S.R.K. Sarma1, M. Ussenaiah2

1K.S.R.K. Sarma, Research Scholar, JNTUA. Regd. No: 13PH0521, Assistant Professor in CSE Department at Vidya Jyothi Institute of Technology (Autonomous), Hyderabad, Telangana, India.
2M. Ussenaiah, Assistant Professor, Dept. of Computer Science, Vikram Simhapuri University, Nellore, Andhra Pradesh, India.
Manuscript received on 07 June 2019 | Revised Manuscript received on 13 June 2019 | Manuscript published on 30 June 2019 | PP: 2185-2191 | Volume-8 Issue-8, June 2019 | Retrieval Number: H7126068819/ 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 classification of texture plays a major role in many image processing applications. This paper proposes an extension to the existing motif co-occurrence matrix (MCM) [1] and its recent variants [2, 3]. This paper initially transforms the color image into HSV color plane and computes the individual color histograms for the H, S and V plane. This paper divides the V-plane of the image into macro regions of size 4×4. Each macro region is divided into four non-overlapped micro regions of size 2×2. Each micro region is replaced with a MCM index which ranges from 0 to 5. This process transforms the macro regions into a grid of size 2×2 with MCM indexes. This paper derives dynamic motif (DM) index on this 2×2 grid and this index ranges from 0 to 23 and extracts region based DM matrix (RDMM) by computing co-occurrence matrix on RDM index image. This paper derives two descriptors based on RDMM and color histogram. The first descriptor computes the histogram on RDMM and integrates these features with the histogram features of H, S and V plane and this form the feature vector. The second descriptor computes the GLCM features on RDMM and integrates with color histograms. The proposed two descriptors are experimented with popular color texture database and the results indicate the efficacy of the proposed method over the existing ones.
Keywords: Dynamic Motifs; Histogram; GLCM Feature; Motif Indexes.

Scope of the Article: Classification