Weighted side-window based gradient guided image filtering

Weimin Yuan, Cai Meng, Xiangzhi Bai

Abstract: Highlights•We propose a weighted-based side window gradient guided filtering (WSGGF), built upon GGF. In WSGGF, both regression and adaptive regularization terms are improved upon side window framework to achieve better edge-preserving performance.•We introduce a more sophisticated minimizing variance-based weighted average strategy (VWA) into our method, which better preserves edges in aggregation process.•We propose a fast version of WSGGF (FWSGGF) that runs about five times faster than the WSGGF while retaining its superior performance on edge-preserving.•We show that the proposed methods achieve superior performances in a variety of applications.