Spatio-temporal super-resolution for multi-videos based on belief propagation
作者：
Highlights：
•
• MAP-MRF model based on Weighted Neighborhood System makes full use of the inter-frame redundant information for robust spatio-temporal super-resolution reconstruction.
• The Belief Propagation algorithm is applied to joint estimate the parameters of MAP-MRF model for edge sharpness and detailed texture preserving, and robustness of noise suppressing.
• The modified SIFT Flow algorithm is robust to the complex spatio-temporal alignment of asynchronous video sequences.
• The proposed MAP-MRF based method produces better objective and subjective performance.
摘要
•MAP-MRF model based on Weighted Neighborhood System makes full use of the inter-frame redundant information for robust spatio-temporal super-resolution reconstruction.•The Belief Propagation algorithm is applied to joint estimate the parameters of MAP-MRF model for edge sharpness and detailed texture preserving, and robustness of noise suppressing.•The modified SIFT Flow algorithm is robust to the complex spatio-temporal alignment of asynchronous video sequences.•The proposed MAP-MRF based method produces better objective and subjective performance.
论文关键词：Super resolution,Spatio-temporal,MAP-MRF,SIFT flow
论文评审过程：Received 5 September 2017, Revised 27 April 2018, Accepted 2 July 2018, Available online 5 July 2018, Version of Record 6 July 2018.
论文官网地址：https://doi.org/10.1016/j.image.2018.07.002
