A Salient Object Detection Method in Images Combining Color and Depth Information

Abstract

The invention discloses an image salient target detection method combining color and depth information. The image enhancement method comprises the following steps: collecting a plurality of color images of a known salient target area and depth images corresponding to the color images; constructing a neural network and inputting the color image and the depth image to respectively obtain a color anddepth information predicted saliency map, setting a branch prediction weight map by the neural network, and obtaining a final fused salient target detection map by the color information predicted saliency map and the depth information predicted saliency map according to the weight map; and constructing a saliency map loss function, a weight map loss function and an edge maintenance loss functionfor training, and processing the to-be-detected color image and the depth image thereof by adopting the trained neural network to obtain a saliency target detection result. According to the method, the most effective parts of the saliency maps predicted by the color and depth information can be selected to be fused, edge details are kept, and a better effect is achieved compared with an image saliency target detection method combining the color and depth information in the past.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to an image salient object detection method combining color and depth information. Background technique [0002] Image salient object detection aims to detect important, eye-catching regions from image scenes that contain a lot of information. Research has found that the human visual mechanism is always able to quickly find and analyze such areas from the scene, while rarely processing the remaining inconspicuous areas. The purpose of the salient target detection task is to construct automated tools to simulate this vision mechanism to extract salient regions in an image. For the subsequent processing, using the salient area as the processing unit can save the time of processing the whole image and improve the processing efficiency of the image. Therefore, salient object detection is widely used in the fields of image segmentation, image retrieval and object r...

AI Technical Summary

Problems solved by technology

[0004] However, how to better integrate color information and depth information is still a problem worth thinking about. The previous methods are still not deep enough for the combination of two different modalities. Only by giving full play to the effective content of color information and depth information , in order to benefit the final test results

Images