Efficient Learned Wavelet Image and Video Coding

Learned wavelet image and video coding approaches provide an explainable framework with a latent space corresponding to a wavelet decomposition. The wavelet image coder iWave++ achieves state-of-the-art performance and has been employed for various compression tasks, including lossy as well as lossless image, video, and medical data compression. However, the approaches suffer from slow decoding speed due to the autoregressive context model used in iWave++. In this paper, we show how a parallelized context model can be integrated into the iWave++ framework. Our experimental results demonstrate a speedup factor of over 350 and 240 for image and video compression, respectively. At the same time, the rate-distortion performance in terms of Bjøntegaard delta bitrate is slightly worse by 1.5% for image coding and 1% for video coding. In addition, we analyze the learned wavelet decomposition by visualizing its subband impulse responses.

1. Introduction

Until the early 2000s, image and video compression based on discrete wavelet transforms has been a highly active field of research. The wavelet transform has desirable properties for compression because its compromise between spatial and frequency resolution fits the correlation structure of images. With the emergence of predictive transform coding as dominant principle in video coding, wavelet-based methods have received less attention. Similarly, the main paradigm in learned image and video compression is nonlinear transform coding [1].