IICNet: A Generic Framework for Reversible Image Conversion

Ka Leong Cheng, Yueqi Xie, Qifeng Chen; Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2021, pp. 1991-2000


Reversible image conversion (RIC) aims to build a reversible transformation between specific visual content (e.g., short videos) and an embedding image, where the original content can be restored from the embedding when necessary. This work develops Invertible Image Conversion Net (IICNet) as a generic solution to various RIC tasks due to its strong capacity and task-independent design. Unlike previous encoder-decoder based methods, IICNet maintains a highly invertible structure based on invertible neural networks (INNs) to better preserve the information during conversion. We use a relation module and a channel squeeze layer to improve the INN nonlinearity to extract cross-image relations and the network flexibility, respectively. Experimental results demonstrate that IICNet outperforms the specifically-designed methods on existing RIC tasks and can generalize well to various newly-explored tasks. With our generic IICNet, we no longer need to hand-engineer task-specific embedding networks for rapidly occurring visual content. Our source codes are available at: https://github.com/felixcheng97/IICNet.

Related Material

[pdf] [arXiv]
@InProceedings{Cheng_2021_ICCV, author = {Cheng, Ka Leong and Xie, Yueqi and Chen, Qifeng}, title = {IICNet: A Generic Framework for Reversible Image Conversion}, booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)}, month = {October}, year = {2021}, pages = {1991-2000} }