A Deep Learning post-processor with a perceptual loss function for video compression artifact removal.
Darren RamsookAnil C. KokaramNeil BirkbeckYeping SuBalu AdsumilliPublished in: PCS (2022)
Keyphrases
- loss function
- video compression
- deep learning
- motion compensation
- motion estimation
- video coding
- unsupervised learning
- video data
- pairwise
- motion compensated
- machine learning
- compression ratio
- support vector
- weakly supervised
- mental models
- image compression
- multi class
- information extraction
- feature extraction
- information retrieval