A Calculation Method of Image Saliency Based on Isoluminance Lines

Abstract

The invention discloses an image significant degree calculation method based on an isolux curve, which comprises the following steps: a to-be-processed image is divided into three image subgraphs in the CIELAB color space and corresponding to an a component, a b component and an L component respectively; potential structure location information and global significance information both of each image subgraph are acquired; a characteristic sub-channel significant image is calculated, and a final image significant image is constructed. Compared with the prior art, the method adopts a mode of integration of global situation and locality to realize extraction of image significant degree information, and non-significant degree information is filtered on the basis of ensuring the accuracy of significant degree information; a non-training mode is adopted to realize extraction of the significant information of the image and any priori or superior information is not required, so that the calculation method is simple and efficient; the same method is adopted to calculate the significant images of three characteristic sub-channels, so that information lose during constructing the final image significant degree image is avoided.

Description

technical field [0001] The invention relates to the technical field of image saliency detection, in particular to an image saliency calculation method based on isoluminance lines. Background technique [0002] The human visual system has always been a magical field for researchers at large. It can process the surrounding complex scenes with amazing efficiency and accuracy, and accurately find the most relevant information in the scene for the current task. People are surrounded by massive amounts of information all the time, which is even far greater than the ability of human brains to process. The human visual system can effectively filter this information, remove useless redundant information and retain the most useful information, and all this is done in an instant. The famous psychologist WilliamJames gave the definition of this characteristic of the human perceptual system as early as 1890: "Everyone knows what attention is. [0003] In computer vision, the detection...

AI Technical Summary

Problems solved by technology

[0007] However, the calculation process of the current method is relatively complicated. If the image needs to be scaled, different algorithms need to be used to extract salient information from different image feature sub-channel maps, and a training process is required; not only that, the extracted image salient information is also easy to loss, susceptibility to complex background and image noise

Images