Exploiting visual descriptions of regions for content-based image retrieval

• Main Author: Tao, Chen.
• Other Authors: Chen, Lihui
• Format: Theses and Dissertations
• Language: English
• Published: 2008
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing; DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems
• Online Access: http://hdl.handle.net/10356/13304
• Institution: Nanyang Technological University

This thesis discusses our view of content-based image retrieval (CBIR) and our experience with an experimental region-of-interest (ROI)-oriented CBIR system. We present a methodology in which efficient representation and indexing of regional visual features serve as the basis for content-based image retrieval. In the experimental system, each image is indexed and accessed by using the features of individual homogeneous regions extracted from the image. Regions in each image are characterized by their visual descriptions, i.e., color and texture properties. The goal of our system is to support the mechanism by which an image may be submitted for regional similarity based query.