Edge Detection Method Based on Multi-Expert Information Fusion

This article presents an information fusion method for edge detection based on multi-expert. There are many methods of edge detection. Each of the methods is regarded as an expert, and every expert has its different detection result because of the different ideas and approaches. It is difficult to determine which result we should choose, because there is no objective standard of choice. PCA is applied to these results. Some components are chosen according to the actual situation. We fuse the information of these chosen components, and derive the single detection result based on multi-expert. This method solves the inconsistency of different experts response. The experiments show the results obtained with this method, both noiseless image and noisy image, and these results are very satisfactory.