Study and Analysis of Liu’s Algorithm for Image Segmentation
Pedapudi Vijaya Bhaskar, Nilamani Bhoi · 2022 Second International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies (ICAECT) · 2022
In this paper the Liu’s algorithm is studied and analyzed. Adaptive control of the range and strength of the iterative pixels is performed here. A novel dissimilarity function is formed to extract the pixels with similar local indices. In the Liu’s method, the over-segmentation which is the first step is performed with mean shift algorithm to keep more number of discontinuities of the image. From this, the method follows HMRF-FCM by modifying its prior probability function and dissimilarity function novelly to include spatial information of the image. A novel prior probability function is formed to distinguish whether a pixel belongs to a homogeneous region or heterogeneous region. The method seems to prevent over smoothing and the results produced on synthetic images, natural images, and radar images imply good segmentation accuracy of the method compared with other state-of-the-art models.