Image retrieval using color features

• Main Author: Revathi Krishnasamy.
• Other Authors: School of Electrical and Electronic Engineering
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems
• Online Access: http://hdl.handle.net/10356/53321
• Institution: Nanyang Technological University

This report aims to present on the work that was undertaken during the one year period of Final Year Project. In this project, a histogram matching based retrieval method is analyzed in three different colour spaces, namely, RGB, HSV and L*a*b. Colour matching is one of the most popular techniques used for retrieving images from a large database but it has drawback whereby information about texture, shape and spatial location are not considered. Colour histogram method requires quantization of a colour space. The quantization process reduces the number of colours used in an image. Two types of quantization: scalar quantization and vector quantization are explored in this project together with the two types of histogram distance measures in three different colour spaces. The extraction algorithm works by: Selecting a colour space(s) {RGB, HSV, L*a*b}, then by selecting a quantization method for the colour space(s), next computation of colour histogram and computation of distance function and lastly, the extraction of images from the database. This report will present the results of an experimental investigation studying on the accuracy, efficiency of the image retrieval system in all three colour models. To test the retrieval performance, Matlab simulation has been implemented. This report also includes the learning points that i have gained during the course of this project as well as how it had enhanced my knowledge throughout the entire project implementation phase.