Premium
8‐2: Training Multi‐Scale Networks for Compression Artifacts Reduction
Author(s) -
Deng Yufan,
Jin Yufeng,
Chen Yinhung,
Zhao Bin
Publication year - 2019
Publication title -
sid symposium digest of technical papers
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.351
H-Index - 44
eISSN - 2168-0159
pISSN - 0097-966X
DOI - 10.1002/sdtp.12860
Subject(s) - reduction (mathematics) , training (meteorology) , computer science , convergence (economics) , compression (physics) , artificial neural network , noise reduction , scale (ratio) , noise (video) , artificial intelligence , data compression , training set , pattern recognition (psychology) , machine learning , algorithm , image (mathematics) , mathematics , geography , materials science , cartography , geometry , meteorology , economics , composite material , economic growth
Present a multi‐scale neural network for compression artifacts reduction. The optimized structure and training procedure can lower the difficulty of convergence and speed up training. Experimental results show that our network outperforms other networks, and can remove complex noise in various sizes, generating cleaner images.