Fast and faithful scale-aware image filters

Abstract This paper proposes a fast and accurate computational framework for scale-aware image filters. Our is based on accurately approximating $$L^{1}$$ L 1 Gaussian convolution with respect to transformed pixel domain representing geodesic distance guidance manifold in order recover salient edges manner faithful scale-space theory while removing small structures. possesses linear complexity high approximation precision. We examined it numerically terms of speed, accuracy, quality compared conventional methods.