Object performance detecting method combined with obviousness information under complicated background

Abstract

The invention discloses an object performance detecting method combined with obviousness information under a complicated background. The method comprises the steps that firstly, obviousness images of all channels under eight set Gaussian kernels are respectively obtained based on the SSS obviousness detecting algorithm; secondly, the Renyi entropy values of the eight obviousness images are respectively calculated, and the largest Renyi entropy value changing position is marked; thirdly, the eight obviousness images are divided into two parts, and the obviousness images with the minimum Renyi entropy values are selected from the two parts respectively; fourthly, the two obviousness images are normalized and overlapped to serve as the final obviousness image of the channel; fifthly, the above steps are repeatedly executed, the respective final obviousness images of the three channels are worked out; sixthly, the obtained final obviousness images of the three channels are multiplied with the respective channel images and combined again; seventhly, the object performance detection is carried out on the combined image through the Bing algorithm. The method has the advantages of being simple in principle, good in operability, high in detecting efficiency and the like.

Description

technical field [0001] The invention mainly relates to the technical field of image processing, in particular to an object property detection method combined with saliency information suitable for complex backgrounds. Background technique [0002] Automatically and quickly detecting objects in images is an important task of computer vision. This research is widely used in automatic image processing, video data analysis of Google glasses, obstacle detection and positioning of automatic robots, etc. [0003] Although great success has been achieved in the field of object recognition, object detection still faces many severe challenges, especially in object detection in 2D images. Because these data do not have depth information, it brings high difficulty to detection. Take one of the most commonly used databases for algorithm evaluation: the VOC2007 database as an example. The database contains 9963 pictures with 20 categories, all of which are common objects in daily life su...

AI Technical Summary

Problems solved by technology

[0005] However, in the existing technology, the way of combining the two is to directly use the segmentation block of the salient area as the candidate object window. The biggest problem of this method is that it depends heavily on the saliency detection algorithm and the segmentation algorithm, and the existing algorithms are very poor. Difficult to meet

Images