Low-rank multi-scale fusion-based image saliency detection method

Abstract

The invention relates to a low-rank multi-scale fusion-based image saliency detection method. The method is technically characterized by comprising the steps of performing single-scale saliency detection on an input image; performing multi-scale saliency fusion processing on the image subjected to the single-scale saliency detection, thereby obtaining a fused saliency map; and performing saliency detailing processing on the fused saliency map after the multi-scale saliency fusion processing to obtain a final cooperative saliency image. According to the method, a low-rank matrix recovery-based saliency detection method and a multi-scale saliency fusion method are applied to the saliency detection, and multi-scale low-rank saliency detection is generalized to cooperative saliency detection of multiple images by applying a GMM-based cooperative saliency priori to detect same or similar regions appearing in the images, so that the problem of difficulty in scale selection is solved, a more reliable saliency detection result is obtained, and the processing capability of the saliency detection is further improved.

Description

technical field [0001] The invention belongs to the technical field of computer vision detection, in particular to an image saliency detection method based on low-rank multi-scale fusion. Background technique [0002] In the field of computer vision, salient object detection methods are divided into two categories: bottom-up scene-driven models and top-down expectation-driven models. Bottom-up methods are mainly based on scene information of picture scenery, while top-down methods are determined by knowledge, expectation and purpose. Many saliency detection methods have been proposed, such as RC, CA, etc. Most of these saliency detection methods are aimed at the saliency detection of single-scale images and have achieved good results. However, these methods have a common problem that when the object is in a natural scene with small scale and high contrast, it is generally unable to detect salient objects in the picture well. For this situation, there are generally two sol...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art, provide an image saliency detection method based on low-rank multi-scale fusion, and solve the problems of difficult scale selection and reliability in the existing detection methods

Images