A Fast Image Retrieval Method Using Multi-Feature Dimension Technique
Kuo Lung Hung, Yih Cheng Tsay · 2009
For querying an image, the most commonly-used method uses low-level features. Of the low-level features, the color feature is the simplest and easiest to understand. Yet, using the color feature alone will not yield a good query result. For example, there may be external factors which cause an alteration in color intensity. To solve these types of problems, in this paper, we propose a multi-feature method composed of saturation, brightness, and frequency to build a three-dimensional space using these features. Then we apply fuzzy theory to color intensity. In this way, our image querying will not fail from an alteration in color intensity. Further experiments have demonstrated that this method is not only effective but is also highly accurate.