This is Matlab implementation of a Bayesian video super-resolution method [1] based on [2].
Compared with the implementation in [2], this implementation is more closer to the details
described in [1], e.g., the update of optical flow.
Description of folders
/data: contains test sequneces, and the results produced by [1]. The original data can be
downloaded from [3].
/celiu_optical_flow: contains optical flow estimation codes, which be downloaded from [4].
Usage
1. Before running, please check and change the path for celiu_optical_flow!
2. run mfsr_cvpr2011_main.m
2.1 city sequence as default, recover the 16th frame in upscale=2 scenario.
2.2 The recovered image is stored in variable TP.Ik
3. You can change the parameters setting in initParam.m, and run mfsr_cvpr2011_main.m.
Different parameters result in different image recover quality.
Just enjoy parameters tunning, and have a fun!
Contact
Peng Qiao, Email: pengqiao@nudt.edu.cn
Citation
[1] C. Liu, and D. Sun, "On Bayesian Adaptive Video Super Resolution," IEEE Trans. on Pattern
Analysis and Machine Intelligence (PAMI), Feb. 2014.
[2] https://github.com/seunghwanyoo/bayesian_vid_sr
[3] http://people.csail.mit.edu/celiu/CVPR2011/default.html
[4] http://people.csail.mit.edu/celiu/OpticalFlow/