Our paper, titled as "Variable bitrate image compression with quality scaling factors", is accepted by the IEEE ICASSP 2020.
Abstract: Recently, learned image compression has emerged with sig- nificant coding efficiency improvement, and even shown no- ticeable gains over the state-of-the-art traditional codecs. In the mean time, most existing methods need to train separate models for different bitrate target. In this paper, we propose to embed a set of quality scaling factors (SFs) into learned image compression network, by which we can encode im- ages across an entire bitrate range with a single model. This solution offers the comparable performance with those de- fault approaches requiring multiple bitrate dependent models, and reduces the complexity significantly for practical imple- mentation. Our work also demonstrates the generalization for various compression network structures, image contents, and training loss functions.
Citation:
Tong Chen, and Zhan Ma, Variable bitrate image compression with quality scaling factors, Proc. of IEEE ICASSP, Barcelona, Spain, May 2020.