Defining a new feature set for content-based image analysis using histogram refinement
✍ Scribed by Jongan Park; Youngan An; Gwangwon Kang; Waqas Rasheed; Seungjin Park; Goorak Kwon
- Publisher
- John Wiley and Sons
- Year
- 2008
- Tongue
- English
- Weight
- 372 KB
- Volume
- 18
- Category
- Article
- ISSN
- 0899-9457
No coin nor oath required. For personal study only.
✦ Synopsis
Abstract
The proposed method is based on color histogram. A new set of features are proposed for content‐based image retrieval (CBIR) in this article. The selection of the features is based on histogram analysis. Standard histograms, because of their efficiency and insensitivity to small changes, are widely used for CBIR. But the main disadvantage of histograms is that many images of different appearances can have similar histograms because histograms provide coarse characterization of an image. We define an algorithm that utilizes the concept of Histogram Refinement (Pass and Zabih, IEEE Workshop on Applications of Computer Vision (1996), 96–102) and we call it color refinement method. Color refinement method splits the pixels in a given bucket into several classes just like histogram refinement method. The classes are all related to colors and are based on color coherence vectors. After the calculation of clusters using color refinement method, inherent features of each of the cluster is calculated. These inherent features include size, mean, variance, major axis length, minor axis length, and angle between x‐axis and major axis of ellipse for various clusters. These inherent features are finally used for image retrieval using Euclidean distance. © 2008 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 18, 86–93, 2008