Saliency map fusion method based on (N) fuzzy integral

Abstract

The invention belongs to the technical field of image processing, and relates to a saliency map fusion method based on the (N) fuzzy integral. Firstly, saliency detection methods to be fused are adopted to generate respective saliency maps. Secondly, the similarity coefficient and similarity matrix between the saliency maps are calculated, and then the support degree and credibility of each saliency map are obtained. Then, the credibility of each saliency map is taken as the fuzzy measure value in the (N) fuzzy integral. At the same time, the saliency maps to be fused are pixel-level ordered,and the sorted discrete saliency values are taken as non-negative real-value measurable functions in the (N) fuzzy integral. Then, the (N) fuzzy integral value is calculated to obtain the fused saliency map, and finally the final saliency map is obtained by enhancing a foreground region and suppressing the optimization of a background region. The method can identify the most salient parts of the images and combine the advantages of the existing excellent saliency detection methods, so that the obtained detection effect is better than that of each synthesis method when the saliency test is performed individually.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to a saliency map fusion method based on (N) fuzzy integral. Background technique [0002] Image saliency detection aims to find the most important part of the image. It is an important preprocessing step in the field of computer vision to reduce computational complexity. It has a wide range of applications in image compression, target recognition, image segmentation and other fields. At the same time, it It is also a challenging problem in computer vision, attracting the research interest of a large number of scholars. At present, a large number of excellent image saliency detection methods have emerged. Each of these methods has its own advantages and disadvantages. Even the same saliency detection method has huge differences in detection effects for different images. Therefore, it is particularly important to be able to fuse the results of multiple saliency detection meth...

AI Technical Summary

Problems solved by technology

There are some traditional saliency map fusion methods, most of which are simple summation or simple multiplication and averaging of multiple saliency maps. The weights of the saliency detection methods are set to the same value, which is unreasonable in practice, because for a picture or even each pixel, the detection effects of various saliency detection methods are different, so each saliency detection method is different. The weights of sex detection methods should also be set differently

There are also some research methods for fusing multiple saliency maps. For example, Mai et al. use conditional random field (CRF) to fuse multiple saliency maps, but the calculation speed is too slow; Qin et al. use multi-layer cellular automata (MCA) To fuse multiple saliency maps, a very good effect is obtained, but the effect of its recall rate is not satisfactory

Images