Image denoising based on spatially adaptive high order total variation model
Lizhen Deng, Qi Fang, Hu Zhu, Gang Zhou · 2016
In order to reduce the staircasing effect of total variation (TV) denoising results, the high order difference is regarded as the regularization item in the high order total variation model (HOTV). The degradation rate of HOTV is faster than the degradation rate of TV in the edge area of the image, so the edge details are lost in the HOTV model while the noise are removed. For reducing the staircasing effect and preserving the edge details effectively, spatially adaptive HOTV model is introduced which difference eigenvalue is as improved adaptive coefficient. The experiment results have demonstrated that the new spatially adaptive HOTV model can preserve detail information and reduce the staircasing effect, at the same time, the de-noising rate and visual effect are superior to both of them.