Generative Compression
Shibani Santurkar, David Budden, Nir Shavit · Picture Coding Symposium · 2018
Traditional image and video compression algorithms rely on hand-crafted encoder/decoder pairs (codecs) that lack adaptability and are agnostic to the data being compressed. We describe the concept of generative compression, the compression of data using generative models, and suggest that it is a direction worth pursuing to produce more accurate and visually pleasing reconstructions at deeper compression levels for both image and video data. We also show that generative compression is orders- of-magnitude more robust to bit errors (e.g., from noisy channels) than traditional variable-length coding schemes.
Cited by 190 works. doi:10.1109/pcs.2018.8456298
Authors
References
- ImageNet classification with deep convolutional neural networks
- Learning Multiple Layers of Features from Tiny Images
- GAN(Generative Adversarial Nets)
- Generative Adversarial Nets
- Auto-Encoding Variational Bayes
- Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network
- Deep Learning Face Attributes in the Wild
- Recognizing human actions: A local SVM approach
- Learning Deep Features for Scene Recognition using Places Database
- Recognizing human actions: a local SVM approach