Super resolution using trilateral filter regression interpolation
Ting-An Chang, Kuan‐Ting Lee, Guan-Cheng Chen, Shu-Hui Chiu, Jar‐Ferr Yang · 2017
The interpolation techniques for super resolution displays and their applications become more and more important since the video programs now are most with high definition. The linear-based interpolation algorithms bring out the jaggy noise noticeably. Recently, the new edge-directed interpolation (NEDI) is proposed to improve the accuracy with one-fold training sizes for predicting parameters. In this paper, the super resolution using trilateral filter regression interpolation (TFRI) is proposed. The trilateral filter is used to modify the adaptive filter suggested in the NEDI method. In the interpolation, the suitable weights are estimated from the pixels in the training window by the proposed trilateral filter. The experimental results demonstrate that the proposed method provides a superior performance such that the edge blurring and blocking effect can be greatly improved in comparison to the other state-of-the-art super resolution methods.