A Hierarchical Nonparametric Discriminant Analysis approach for Content-Based Image Retrieval system. (11th September 2006)

• Record Type: Journal Article
• Authors: Chung, Kien-Ping; Fung, Chun Che
• Is Part Of: International journal of electronic business. Volume 4:Number 3/4(2006)
• Page Start: 208
• Page End: 220
• Publication Date: 2006-09-11
• Subjects: content-based image retrieval -- CBIR -- image retrieval -- hierarchical nonparametric discriminant analysis -- HNDA -- relevance feedback -- RF -- feature selection -- electronic business -- e-business -- business integration; Electronic commerce -- Periodicals; 658.05467805
• Languages: English
• ISSNs: 1470-6067

Abstract

This paper proposes a Hierarchical Nonparametric Discriminant Analysis (HNDA) content-based image retrieval (CBIR) system for e-business applications. Developments in CBIR have drawn much interest in recent years. The challenge is how to retrieve the most appropriate or relevant images at the fastest speed. In this paper, a hierarchical multi-layer statistical discriminant framework is proposed. The system is able to select the most appropriate features by analysing the newly received images, and then apply a Relevance Feedback (RF) approach to improve the retrieval accuracy. As the number of features being analysed is less, an improvement in performance is achieved.