C++ implementation of the Fast Super-Resolution Convolutional Neural Network (FSRCNN).
This implements two models: FSRCNN which is more accurate but slower and FSRCNN-s which is faster but less accurate.
Based on this paper.
- >Eigen-3.3.7
- >OpenCV-3.4.5
Usage: FSRCNN-OpenCV.exe <option(s)> SOURCE-IMG
Options:
-f,--fast True/False Use FSRCNN-s or FSRCNN
-s,--scale 2/3 Specify the scale num
Original Comic image:
Bicubic Comic image X3:
FSRCNN Comic image X3: