- Tutor: Bart Selman from Cornell University
- Purpose: Used symmetric padding to improve perceptual quality of image after super-resolution
- Duties: Edited code to realize VDSR using tensorflow and symmetric padding to process the images with three channels; Used EC2 of AWS to do experiments using classic Datasets (ImageNet, Set5, Set14) and evaluated the outcomes via PSNR and SSIM; Wrote the Introduction and Experiment&Analysis of the final paper
- Final paper and Recommendation Letter (<-CLICK)