A robust simple linear iterative clustering algorithm
Shiren Li, Junwei Huang, Jiayu Shang, Xiongyi Wei · 2017
This paper presents a novel superpixel algorithm - a robust simple linear iterative clustering (R_SLIC) algorithm. The traditional SLIC algorithm is allergic to noise. In order to increase the robustness of the algorithm and make it noise-insensitive, we modify the framework of SLIC by supplementing a regularized term. The regularized term can balance the effect of various parts. What's more, the initial step of assigning seeds is also modified to avoid locating the center on an edge and reduce the opportunity of choosing a noisy pixel as the center. From the experimental results, it is demonstrated that the proposed algorithm is fast and robust to noise.