High-Fidelity Image Compression with Score-based Generative Models.
Emiel HoogeboomEirikur AgustssonFabian MentzerLuca VersariGeorge TodericiLucas TheisPublished in: CoRR (2023)
Keyphrases
- high fidelity
- generative model
- image compression
- probabilistic model
- vector quantization
- wavelet transform
- mixture model
- real time
- discriminative models
- conditional random fields
- compression algorithm
- prior knowledge
- semi supervised
- em algorithm
- discriminative learning
- high quality
- compression ratio
- high resolution
- hierarchical hidden markov models
- image processing
- generative and discriminative models
- video data
- human operators
- training set
- object recognition
- computer vision