arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2007.05146cs.CV
收藏

光流蒸馏:实现高效稳定的视频风格传输

Optical Flow Distillation: Towards Efficient and Stable Video Style Transfer

Noah’s Ark Lab, Huawei Technologies(华为技术有限公司诺亚方舟实验室);University of Sydney(悉尼大学)

Xinghao Chen, Yiman Zhang, Yunhe Wang, Han Shu, Chunjing Xu, Chang Xu

摘要:

视频风格传输技术激发了移动设备上许多令人兴奋的应用。然而,它们的效率和稳定性仍然远远不能令人满意。为了提高跨帧的传输稳定性,尽管光流具有高计算复杂性(例如,占用超过97%的推理时间),但其被广泛采用。提出了一种基于知识蒸馏范式的轻量级视频风格传输网络。我们采用两个教师网络,其中一个采取光流在推理过程中,而另一个没有。这两个教师网络之间的输出差异突出了光流所做的改进,然后采用光流来提取目标学生网络。此外,低秩蒸馏损失被用来稳定学生网络的输出,通过模仿输入视频的秩。大量的实验表明,我们的学生网络没有光流模块仍然能够生成稳定的视频,运行速度比教师网络快得多。

英文摘要:

Video style transfer techniques inspire many exciting applications on mobile devices. However, their efficiency and stability are still far from satisfactory. To boost the transfer stability across frames, optical flow is widely adopted, despite its high computational complexity, e.g. occupying over 97% inference time. This paper proposes to learn a lightweight video style transfer network via knowledge distillation paradigm. We adopt two teacher networks, one of which takes optical flow during inference while the other does not. The output difference between these two teacher networks highlights the improvements made by optical flow, which is then adopted to distill the target student network. Furthermore, a low-rank distillation loss is employed to stabilize the output of student network by mimicking the rank of input videos. Extensive experiments demonstrate that our student network without an optical flow module is still able to generate stable video and runs much faster than the teacher network.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑