Image retrieval using color and shape features

• Main Author: Foram Kothari.
• Other Authors: Tan Yap Peng
• Format: Theses and Dissertations
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing
• Online Access: http://hdl.handle.net/10356/18811
• Institution: Nanyang Technological University

Technological advances in digital imaging and broadband networking have facilitated millions of ordinary people to express themselves and communicate with others by sharing images, video, and other forms of media online. This necessitates development of effective and efficient tools to index and search multimedia information. A technique for Content-Based Image Retrieval (CBIR) is presented to conduct similarity queries in image databases, which are based on some visual features of images, such as color, texture, shape, and regions. This thesis attempts to understand some issues in CBJR and implement a simple, robust and effective image retrieval algorithm. Important concepts such as feature extraction, similarity measures, database indexing and retrieval efficiency are discussed. The image retrieval approach studied is based on Query by Visual Example. A directional distance histogram is reviewed and a modified approach is proposed which is rotation invariant and more robust. A retrieval scheme based on color histogram for HSI color space is also implemented which has a very good retrieval performance. Finally a fused algorithm is developed which integrates the above two approaches for image retrieval.