Grid-search wolf pack optimization algorithm for two-dimensional OTSU image segmentation

Abstract

Compared with One-Dimensional OTSU, Two-Dimensional OTSU includes the relationship about the pixels of the image and improves the anti-noise ability of the algorithm. However, as a result of too much computation caused by considering the relationship about pixels of the image to be processed, Two-Dimensional OTSU always takes much time. To keep the good performance of original Two-Dimensional OTSU as well as reduce the time spent; this paper proposes Grid-Search Wolf Pack Optimization Algorithm (GSWA) for Two-Dimensional OTSU Image Segmentation. Firstly, it is based on the thought of adaptive shrinking grid search chaos wolf optimization algorithm with adaptive standard-deviation updating amount (ASGS-CWOA) that traditional wolf pack optimization algorithm was improved to enhance its performance, which includes Grid Search and Opposite-Middle Raid(OMR). Grid Search can enhance the wolf pack’s summon and siege capability. Moreover, discrete step-size was adopted to adapt to discrete solution space of Image-Segmentation. Experimental results indicate that GSWA-OTSU not only reduces the segmentation time, but also takes the segmentation quality into account.