SC-PROSAC: An Improved Progressive Sample Consensus Algorithm Based on Spectral Clustering

作者：Shuhua Ma, Xianchun Ma, Hairong You, Tiang Tang, Jinhua Wang, Min Wang · 年份：2021 · DOI：10.1109/icrcv52986.2021.9546964 · 被引用次数：5 · 研究领域：Advanced Computing and Algorithms、Retinal Imaging and Analysis、Brain Tumor Detection and Classification

This paper proposes an improved progressive consensus sampling (PROSAC) algorithm based on the spectral clustering algorithm, which aiming at the problem that the calculation model of PROSAC may fail, resulting in low final matching accuracy. First, the spectral clustering algorithm is used to filter the generated subset of the PROSAC algorithm, and then based on the selected subset, a high-precision calculation model is generated, and finally the final match is obtained according to the calculation model. The experimental results show that the calculation model generated by the improved PROSAC algorithm has high accuracy, thereby improving the accuracy and efficiency of image matching.