ANFIC: Image Compression Using Augmented Normalizing Flows

• Published in: IEEE open journal of circuits and systems Vol. 2; pp. 613 - 626
• Main Authors: Ho, Yung-Han, Chan, Chih-Chun, Peng, Wen-Hsiao, Hang, Hsueh-Ming, Domanski, Marek
• Format: Journal Article
• Language: English
• Published: New York IEEE 2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: augmented normalizing flows; Codes; Convolutional neural networks; Decoding; Degradation; flow-based image compression; Image coding; Image compression; Learning-based image compression; Machine learning; Network architecture; perceptually lossless image compression; Source code; Stacking; Training data; variable rate image compression
• ISSN: 2644-1225; 2644-1225
• DOI: 10.1109/OJCAS.2021.3123201

This paper introduces an end-to-end learned image compression system, termed ANFIC, based on Augmented Normalizing Flows (ANF). ANF is a new type of flow model, which stacks multiple variational autoencoders (VAE) for greater model expressiveness. The VAE-based image compression has gone mainstream, showing promising compression performance. Our work presents the first attempt to leverage VAE-based compression in a flow-based framework. ANFIC advances further compression efficiency by stacking and extending hierarchically multiple VAE's. The invertibility of ANF, together with our training strategies, enables ANFIC to support a wide range of quality levels without changing the encoding and decoding networks. Extensive experimental results show that in terms of PSNR-RGB, ANFIC performs comparably to or better than the state-of-the-art learned image compression. Moreover, it performs close to VVC intra coding, from low-rate compression up to perceptually lossless compression. In particular, ANFIC achieves the state-of-the-art performance, when extended with conditional convolution for variable rate compression with a single model. The source code of ANFIC can be found at https://github.com/dororojames/ANFIC .