SLIC super-pixel block optimizing method

Abstract

The invention discloses an SLIC super-pixel block optimizing method. The method includes the steps that (1) initialized super-pixel blocks are obtained; (2) initialized super-pixel block seeds are obtained; (3) the super-pixel blocks on each line serve as target zones, a search range is determined, the number of iteration times is designated, and update processing is conducted, wherein (3-1) the search range is selected according to target line seeds and pixel point distance measuring factor values are calculated; (3-2) in the overlapped search range of the current super-pixel block seeds and other super-pixel block seeds, the pixel point distance measuring factor values are compared, and pixel points closer to the current super-pixel block seeds are clustered into the current super-pixel blocks; (3-3) the updated super-pixel blocks are obtained, updated super-pixel block seeds on target lines are calculated, whether the number of iteration times is zero or not is judged, if not, the step (3-1) is executed again according to the updated target lines, and if yes, the finally-updated super-pixel blocks are obtained. The method has the advantages of being small in calculated amount and convenient to conduct through hardware.

Description

technical field [0001] The invention belongs to the field of image segmentation, in particular to a SLIC superpixel block optimization method. Background technique [0002] Image segmentation refers to dividing an image into different regions with special semantics according to a certain similarity criterion, so as to separate the target of interest in a complex background environment, which is a basic problem in image analysis, pattern recognition and computer vision. Superpixel segmentation is mainly used in pre-segmentation in image segmentation. Superpixels refer to image blocks composed of adjacent pixels with similar texture, color, brightness and other characteristics. He uses the similarity of features between pixels to group pixels. The redundant information of the image can be obtained, and the redundant information of the image after superpixel segmentation is less, and the subsequent processing is performed with superpixels as a unit, which greatly reduces the co...

AI Technical Summary

Problems solved by technology

However, when considering the hardware implementation of the above algorithm, the cost must be considered, and the calculation amount of the SLIC superpixel segmentation algorithm determines the hardware cost, and the seeds of the traditional SLIC superpixel segmentation method will float and the position is unstable

Images