Evolvement Constrained Adversarial Learning for Video Style Transfer
This work improves video style transfer for applications like augmented reality and games, but it is incremental as it builds on existing adversarial learning frameworks.
The paper tackles video style transfer by introducing an evolve-sync loss to replace optical flow, addressing sensitivity to occlusions and rapid motions, and shows quantitative and qualitative improvements over state-of-the-art methods.
Video style transfer is a useful component for applications such as augmented reality, non-photorealistic rendering, and interactive games. Many existing methods use optical flow to preserve the temporal smoothness of the synthesized video. However, the estimation of optical flow is sensitive to occlusions and rapid motions. Thus, in this work, we introduce a novel evolve-sync loss computed by evolvements to replace optical flow. Using this evolve-sync loss, we build an adversarial learning framework, termed as Video Style Transfer Generative Adversarial Network (VST-GAN), which improves upon the MGAN method for image style transfer for more efficient video style transfer. We perform extensive experimental evaluations of our method and show quantitative and qualitative improvements over the state-of-the-art methods.