A Convolutional Neural Network-Based Saliency Detection Method Based on Region- and Pixel-Level Fusion

Abstract

The invention discloses a saliency detection method based on convolutional neural network region and pixel-level fusion. The research object of this method is a static image, wherein the content of the image can be arbitrary, and the research goal is to find out Objects that draw the eye's attention and assign them different salience values. The present invention mainly proposes an adaptive region generation technology, and designs two CNN network structures, which are respectively used for pixel-level saliency prediction and saliency fusion. These two CNN network models take images as input, use the real results of images as supervisory signals for the training of network models, and finally output a saliency map with the same size as the input image. The present invention can effectively perform region-level saliency estimation and pixel-level saliency prediction, obtain two saliency maps, and finally use CNN for saliency fusion to fuse the two saliency maps and the original image to obtain the final saliency map .

Description

technical field [0001] The invention relates to an image processing method based on deep learning, in particular to a saliency detection method based on convolutional neural network region and pixel-level fusion. Background technique [0002] With the development and rise of deep learning, the saliency detection technology based on deep learning is also developing continuously. Saliency detection can be divided into two categories: bottom-up data-driven models and top-down task-driven models. Bottom-up saliency detection refers to finding an eye-catching object in a given image, which can be any type of thing. The top-down saliency detection method usually finds objects of a given category from a given picture and assigns different saliency values. Currently, bottom-up saliency detection methods are the most studied. [0003] Existing bottom-up saliency detection methods can be divided into two categories, methods based on manually designed features and methods based on c...

AI Technical Summary

Problems solved by technology

However, these methods cannot obtain accurate pixel-level saliency prediction results.

Images