Performance Evaluation on Image Segmentation based on FD-weighed Grey Relational Analysis

It is important for the design and the modification of segmentation algorithms to acquire more reasonable evaluation results on image segmentation. The paper proposes an efficient evaluation method, which is based on FD-weighted grey relational analysis. In the method, as far as specific purpose, application circumstance and research interest are concerned, performance indicators are selected first. And then, their relative importance is appointed via Forced Decision method. Finally, segmentation performance are analyzed and sorted on weighted grey relational degrees. Experimental results show that the method is efficient, fast and stable for providing with an overall performance.