Lossy Image Compression with Quantized Hierarchical VAEs
This work addresses efficient image compression for applications requiring high-quality, fast processing, representing an incremental improvement over existing VAE-based methods.
The paper tackles lossy image compression by redesigning ResNet VAEs with quantization-aware latent models and improved architecture, achieving superior performance over previous methods on natural images and enabling fast GPU execution.
Recent research has shown a strong theoretical connection between variational autoencoders (VAEs) and the rate-distortion theory. Motivated by this, we consider the problem of lossy image compression from the perspective of generative modeling. Starting with ResNet VAEs, which are originally designed for data (image) distribution modeling, we redesign their latent variable model using a quantization-aware posterior and prior, enabling easy quantization and entropy coding at test time. Along with improved neural network architecture, we present a powerful and efficient model that outperforms previous methods on natural image lossy compression. Our model compresses images in a coarse-to-fine fashion and supports parallel encoding and decoding, leading to fast execution on GPUs. Code is available at https://github.com/duanzhiihao/lossy-vae.