Abstract

In this paper, we present a new and effective method for detecting regions matching a given color, along with the features of the region surfaces. We adopt the HVC color coordinates in our method because of its ability to completely separate the luminant and chromatic components of colors. A second-order basis function is introduced to pick up the pixels matching the specified color features. The convolution operation involving the basis function is incorporated to diminish the influence of noise and increase the integrity of the detected regions. Efforts are made to remarkably reduce the running time for successive region detection of the same image. Various experimental results show the effectiveness of the proposed method in handling images under shade, highlight, sharp contrast, and other nonuniform illumination conditions.