Colour feature based analysis on content-based image retrieval techniques

• Main Authors: Tuan Muda, Tuan Zalizam, Din, Roshidi, Kamaruddin, Siti Sakira, Yusof, Yuhanis
• Format: Article
• Language: English
• Published: AICIT, Korea 2014
• Subjects: QA76 Computer software
• Online Access: http://repo.uum.edu.my/15360/1/464d44.pdf; http://repo.uum.edu.my/15360/; http://www.aicit.org/jdcta/home/index.html

Content-based image retrieval is an active research area that has been famous since 1990s. The main objective of content-based image retrieval is to present users with the most similar images to the search query.This study discusses on the comparative analysis of three techniques frequently used in CBIR which are GLCM, K-Means, and Gabor filtering.The undertaken work focuses on utilizing colour histogram to represent stored images as well as the query. A set of 1000 images that are of 6 categories that includes bus, beach, building, flower, horse, and mountain are used in the analysis.Similarity between an image query and the stored images is calculated based on the Euclidean distance measure. Based on the undertaken experiments, it is learned that the GLCM retrieval model produces a better result as compared to the K-Means and Gabor filtering models.