Image Segmentation Based on Block Level and Hybrid Directional Local Extrema

• DOI: 10.32604/cmc.2022.018423
• Publication Date: 2021-09-27T02:49:01Z
• Authors: Ghanshyam Raghuwa...; Yogesh Gupta; Deepak Sinwar; Dilbag Singh; Usman Tariq; Muhammad Attique; Kuntha Pin; Yunyoung Nam
• Keywords: Range segmentation; Segmentation-based object categorization; Region growing; Maxima and minima

In the recent decade, digitalization of various tasks has added great flexibility to human lifestyle and changed daily routine activities communities. Image segmentation is a key step in digitalization. Segmentation plays role almost all areas image processing, approaches have been proposed for segmentation. this paper, novel approach using nonuniform adaptive strategy. Region-based along with directional binary pattern generated better segmented image. An mask 8 × was circulated over pixels whose bit value 1 pattern. performed three phases: first, an divided into sub-images or chunks; next, patches were taken as input, threshold generated; finally chunks processed separately by convolving on chunks. Gradient Laplacian Gaussian algorithms extrema patterns provided double check boundary pixels. The tested varying sizes, after multiple iterations, it found that block size performs than other sizes. accuracy technique measured terms count ill regions, which extracted process.