Lossy Image Compression with Recurrent Neural Networks: from Human Perceived Visual Quality to Classification Accuracy.
Maurice WeberCédric RenggliHelmut GrabnerCe ZhangPublished in: CoRR (2019)
Keyphrases
- visual quality
- recurrent neural networks
- image quality
- bit rate
- high frequency
- human visual system
- high compression ratio
- reconstructed image
- neural network
- feed forward
- error concealment
- human observers
- compressed images
- low bit rate
- artificial neural networks
- visual information
- peak signal to noise ratio
- feature selection
- quality assessment
- high quality
- feature space
- image compression
- computational complexity
- image data