| Abstract: | This paper aimed at segmentation of natural images, in which the color and texture of each segment does not typically exhibit uniform statistical characteristics due to effect of lighting, perspective, scale changes, etc. Although significant progress has been made in texture segmentation and color segmentation separately, the area of combined color and texture segmentation remains open and active. The proposed approach is based on two types of spatially adaptive low-level features. The first describes a spatially adaptive dominant color, obtained using Adaptive Clustering Algorithm (ACA), and second is the spatial characteristic of the grayscale component of the texture, obtained from the detail coefficient of Complex Wavelet Decomposition (CWD). Together, they provide a simple and effective characterization of texture that the proposed algorithm uses to obtain robust and, at the same time, accurate and precise segmentations. Initially, image pixels are classified into smooth and non-smooth or texture pixels. The smooth pixels are segmented using the morphological watershed algorithm. The non-smooth pixels are segmented by two steps. First step is the crude segmentation by Multi-grid Region Growing algorithm and the second step is the iterative edge refinement. The resulting segmentations convey semantic information that can be used for content-based retrieval. The performance of the proposed algorithms is numerically evaluated, compared with other algorithms and found to be the best.
|