C-STCM: Image Retrieval using Color Based Stretched Text on Co-Occurrence Matrix

  • M. Vijayashanthi, Dr. G. Venkata Rami Reddy

Abstract

In this paper, a novel feature descriptor named stretched texton co-occurrence matrix (STCM) is “proposed for content-based image retrieval (CBIR)”. The STCM represents the shape, color and texture features more efficiently and precisely. The color information is extracted by quantizing each component of HSV color space. The shape features in the form of stretched texton indexes (STI) are computed by quantizing the V component of HSV color space. The earlier texton-based approaches derived only a few textons on a microgrid. The proposed method extracted all probable textons on a 2x2 grid and considered the magnitude relationship among the pixels that are part and not part of texton formation. This aspect is not considered in the earlier methods. The co-occurrence features are obtained by transforming the identified microstructures in the form of STI image into STCM. This paper integrated the color features derived from the histograms of individual HSV color space, with new structural patterns “stretched Texton’s”, texture features in the form of GLCM and this process derived the proposed Color-Stretched Texton Co-occurrence Matrix (C-STCM) .  The experiments are performed on natural databases, i.e., Corel- 1K, Corel 10K, gray and color Broadtaz texture, CMU-PIR, and MIT-VisTex. The results obtained depict the outstanding performance of the proposed C-STCM over the state-of-art methods.

Published
2021-11-22
How to Cite
M. Vijayashanthi, Dr. G. Venkata Rami Reddy. (2021). C-STCM: Image Retrieval using Color Based Stretched Text on Co-Occurrence Matrix. Design Engineering, 14412 - 14435. Retrieved from http://thedesignengineering.com/index.php/DE/article/view/6571
Section
Articles