Multi--Scale Recursive and Perception--Distortion Controllable Image Super--Resolution

Pablo Navarrete Michelini, Dan Zhu, Hanwen Liu; The European Conference on Computer Vision (ECCV) Workshops, 2018, pp. 0-0

Abstract


We describe our solution for the PIRM Super–Resolution Challenge 2018 where we achieved the 2nd best perceptual quality for average RMSE 􏰀 16, 5th best for RMSE 􏰀 12.5, and 7th best for RMSE 􏰀 11.5. We modify a recently proposed Multi–Grid Back– Projection (MGBP) architecture to work as a generative system with an input parameter that can control the amount of artificial details in the output. We propose a discriminator for adversarial training with the following novel properties: it is multi–scale that resembles a progressive– GAN; it is recursive that balances the architecture of the generator; and it includes a new layer to capture significant statistics of natural images. Finally, we propose a training strategy that avoids conflicts between reconstruction and perceptual losses. Our configuration uses only 281k parameters and upscales each image of the competition in 0.2s in average.

Related Material


[pdf]
[bibtex]
@InProceedings{Michelini_2018_ECCV_Workshops,
author = {Navarrete Michelini, Pablo and Zhu, Dan and Liu, Hanwen},
title = {Multi--Scale Recursive and Perception--Distortion Controllable Image Super--Resolution},
booktitle = {The European Conference on Computer Vision (ECCV) Workshops},
month = {September},
year = {2018}
}