/11-785-Project

11-785 Project

Primary LanguagePython

11-785-Project

Dataset

Urban 100 from "Single Image Super-Resolution from Transformed Self-Exemplars"

PDF: https://ieeexplore.ieee.org/document/7299156

Baseline Model 0: SRCNN

"Image Super-Resolution Using Deep Convolutional Networks"

PDF: https://ieeexplore.ieee.org/abstract/document/7115171/

Model 1: FSRCNN

"Accelerating the super-resolution convolutional neuralnetwork"

PDF: http://arxiv.org/abs/1608.00367

Model 2: SRResNet

"Photo-realistic single image super-resolution using a generative adversarial network"

PDF: http://arxiv.org/abs/1609.04802

Model 3: CAR-variant

Inspired by

"Learned image downscaling for upscaling using content adaptive resampler"

PDF: https://arxiv.org/pdf/1907.12904.pdf

RUN

SRCNN python run.py --bicubic 1

FSRCNN python run.py --model FSRCNN --bicubic 1 python run.py --model FSRCNN --bicubic 0

SRResNet python run.py --model SRResNet --bicubic 1 python run.py --model SRResNet --bicubic 0

CAR-variant python run.py --model car --bicubic 1 python run.py --model car --bicubic 0

where bicubic = 1 indicates using the interpolated dataset and bicubic = 0 indicates using the original Urban100 dataset.