Neural Multi-scale Image Compression

This study presents a new lossy image compression method that utilizes the multi-scale features of natural images. Our model consists of two networks: multi-scale lossy autoencoder and parallel multiscale lossless coder. The multi-scale lossy autoencoder extracts the multiscale image features to quantized variables, and the parallel multi-scale lossless coder enables rapid and accurate lossless coding of the quantized variables via encoding/decoding the variables in parallel. Our proposed model achieves comparable performance to the state-of-the-art model on Kodak and RAISE-1k dataset images, and it encodes a PNG image of size 768x512 in 70 ms with a single GPU and a single CPU process and decodes it into a high-fidelity image in approximately 200 ms.