Abstract | ||
---|---|---|
This paper presents a novel approach for image region segmentation based on color coherence quantization. Firstly, we conduct an unequal color quantization in the HSI color space to generate representative colors, each of which is used to identify coherent regions in an image. Next, all pixels are labeled with the values of their representative colors to transform the original image into a "Color Coherence Quantization" (CCQ) Labels with the same color value are then viewed as coherent regions in the CCQ image. A concept of "connectivity factor" is thus defined to describe the coherence of those regions. Afterwards, we propose an iterative image segmentation algorithm by evaluating the "connectivity factor" distribution in the resulted CCQ image, which results in a segmented image with only a few important color labels. Image segmentation experiments of the proposed algorithm are designed and implemented on the MSRC datasets [1] in order to evaluate its performance. Quantitative results and qualitative analysis are finally provided to demonstrate the efficiency and effectiveness of the proposed approach. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1007/978-3-642-21822-4_22 | IEA/AIE (1) |
Keywords | Field | DocType |
image segmentation experiment,color coherence quantization,image region segmentation,connectivity factor,segmented image,iterative image segmentation algorithm,coherent region,representative color,ccq image,original image,color quantization,qualitative analysis,image segmentation,color space | Computer vision,Scale-space segmentation,Color histogram,Computer science,Image texture,Binary image,Image segmentation,Color balance,Artificial intelligence,Color quantization,Color image | Conference |
Citations | PageRank | References |
1 | 0.36 | 9 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Guang-Nan He | 1 | 5 | 1.41 |
Yubin Yang | 2 | 277 | 33.46 |
Yao Zhang | 3 | 21 | 4.52 |
Yang Gao | 4 | 263 | 58.38 |
Lin Shang | 5 | 526 | 31.36 |