python_iccv15.zip_GT3_image

# Python demo code for image super-resolution using sparse coding network By Zhaowen Wang (zhawang@adobe.com), 2015 This code is for academic purpose only. Not for commercial/industrial activities. # Required libraries * Python 2.7 * OpenCV 2.4 * Matlab/pymatbridge (only for generating low-res images) # Models - 52: x2, 128 dictionary size, LISTA structure, no recurrent structure, SHLU activation - 310: x2, cascade network for beyond x2 SR # Note Minor difference from the results reported in the ICCV paper may exist due to implementation details. # Citation @inproceedings{wang2015deep, Author = {Wang, Zhaowen and Liu, Ding and Yang, Jianchao and Han, Wei and Huang, Thomas S}, Title = {Deep Networks for Image Super-Resolution with Sparse Prior}, Booktitle = {IEEE International Conference on Computer Vision (ICCV)}, Year = {2015}, Organization = {IEEE} } Paper link: http://arxiv.org/abs/1507.08905