Deeply-Recursive Convolutional Network for Image Super-Resolution
Jiwon KimDepartment of ECE, Seoul National University, KoreaJung Kwon LeeDepartment of ECE, Seoul National University, KoreaKyoung Mu LeeDepartment of ECE, Seoul National University, Korea
2016en
ABI
Abstract
We propose an image super-resolution method (SR) using a deeply-recursive convolutional network (DRCN). Our network has a very deep recursive layer (up to 16 recursions). Increasing recursion depth can improve performance without introducing new parameters for additional convolutions. Albeit advantages, learning a DRCN is very hard with a standard gradient descent method due to exploding/ vanishing gradients. To ease the difficulty of training, we propose two extensions: recursive-supervision and skip-connection. Our method outperforms previous methods by a large margin.
Identifiers
Citations and references
Cited by 30 references