The improved local and global contrast image saliency detection algorithm
Fei Yao · Academic Journal of Computing & Information Science · 2020
Contrast description is a very important and commonly used feature description method in the image saliency detection, according to the difference of contrastive region, the detection algorithm can be divided into local and global contrast algorithm. In this paper, firstly, the local contrast AC algorithm is improved, from two aspects of distance calculation method, weighting coefficient and multi-scale average, to improve the problem of emphasizing small and medium-sized areas of significant targets in the original AC algorithm, the experimental results show the effectiveness of the improved scheme. Then the global contrast RC algorithm is analyzed and improved, and the original RC algorithm is improved from two aspects of image pre-segmentation Algorithm improvement and significant center position optimization, the experimental results show the effectiveness of the improved scheme.