We ranked 6th at LG-Dacon Challenge
2021.08.31
Our Lab participated in LG-Dacon Image Quality Improvement challenge.
- π Leaderboard
The main task of this competition is image quality improvement. The participants restore the input image corrupted by light smearing to target image.
They proposed FiLM square, an architecture based on Fully Convolutional U-Net network and FiLM(Feature-wise Linear Modulation). After the presentation they finally ranked 6th at that competition.
All the proposed models will be fully open-sourced.
Congrats!