Texture Classification Based On Fuzzy Similarity Texton Co-Occurrence Matrix
J.Srinivas1, Ahmed Abdul Moiz Qyser2, B. Eswara Reddy3
1J.Srinivas, Research Scholar, Muffakham Jah College of Engineering and Technology, Hyderabad, Telangana, India.
2Dr. Ahmed Abdul Moiz Qyser, Department of Computer Science and Engineering, Muffakham Jah College of Engineering and Technology, Hyderabad, Telangana, India.
3Dr. B. Eswara Reddy, Principal, JNTUA College of Engineering, Kalikiri Chittoor, Andhra Pradesh, India.
Manuscript received on 10 December 2018 | Revised Manuscript received on 17 December 2018 | Manuscript Published on 30 December 2018 | PP: 455-463 | Volume-8 Issue- 2S December 2018 | Retrieval Number: BS2723128218/19©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: In the existing texton based methods a texton is derived in a grid by a collection of pixels exhibiting exactly the similar grey level values/color/attributes. The disadvantage of this approach is they fail in recognizing textons, whenever a small random noise changes the pixels intensity values slightly. This paper addresses this by deriving a fuzzy similarity ‘S’ in identification of texton patterns. The proposed Fuzzy similarity Texton Co-occurrence Matrix (FSTCM) framework considers the pixels whose gray level value falls within the fuzzy similarity index value as texton pattern. The FSTCM divides initially the texture image into micro regions of size 2×2, identifies the textons and transforms the texture image into a fuzzy texton image. This paper derives gray level co-occurrence matrix (GLCM) features on FSTCM and the proposed method is tested on five popular texture image databases. The experimental investigation reveals the high performance of the proposed method over the state of art local based and texton based methods.
Keywords: Texton, Similarity; Micro Region; GLCM Features; Random Noise.
Scope of the Article: Classification