Multi-mapping Image-to-Image Translation via Learning Disentanglement

However, the existing methods only consider one of the two perspectives, which makes them unable to solve each other's problem. To address this issue, we propose a novel unified model, which bridges these two objectives.