A Scalable Content-Based Image Retrieval System Using Color Coherence Vectors and Wavelet Decomposition

International Journal of Computer Science and Technology | Vol. 13, No. 1, January 2011 | pp. 1–7
DOI: 10.46882/2011/IJCST/000107

Research Article

Title: A Scalable Content-Based Image Retrieval System Using Color Coherence Vectors and Wavelet Decomposition

Names of Authors: Li Wei¹, Francois Dubois²

Authors’ Affiliations: ¹Department of Computer Science, Tsinghua University, Beijing, China; ²Laboratoire d'Informatique, Sorbonne Université, Paris, France

Abstract: The exponential growth of multimedia repositories demands highly scalable Content-Based Image Retrieval (CBIR) methods capable of extracting relevant images with minimal computational complexity. Conventional systems based solely on global color histograms lose critical spatial data, returning high false matches. This paper presents an advanced CBIR architecture combining Discrete Wavelet Transform (DWT) and Color Coherence Vectors (CCV). The system applies a two-level Haar DWT to decouple structural textures into sub-bands, while spatial distributions of color profiles are derived via CCVs to partition pixels into coherent and incoherent clusters. Multi-dimensional indexing is optimized through an R-tree structural model to accelerate database query execution times. Experiments were performed on the Corel 1,000-image benchmark collection. The combined DWT-CCV approach yielded a mean precision rate of 84.6% and a recall rate of 79.2%, outperforming standard histogram-based approaches by over 18.5%. The query execution runtime averaged 124 milliseconds per search on a standard database setup, establishing high scalability characteristics. This framework provides an industrial-grade solution for digital asset management, stock photo licensing platforms, and online product catalog search systems.

Keywords: Content-Based Image Retrieval, Discrete Wavelet Transform, Color Coherence Vector, Feature Extraction, Computer Vision

Manuscript Timeline: Received August 20, 2010; Revised September 25, 2010; Accepted October 06, 2010; Published January 18, 2011.

Citation: Wei, L., & Dubois, F. (2011). A Scalable Content-Based Image Retrieval System Using Color Coherence Vectors and Wavelet Decomposition. International Journal of Computer Science and Technology, 13(1), 1–7.

Subscribe to read the full article: https://www.internationalscholarsjournals.org/subscribe-to-read