Optical Flow Regularization of Implicit Neural Representations for Video Frame Interpolation

Weihao Zhuang, Tristan Hascoet, Ryoichi Takashima, Tetsuya Takiguchi

Keywords: Implicit Neural Representation, Video Representation, Video Frame Interpretation

TL;DR: We show that constraining the derivatives of video INR to satisfy the optical flow constraint equation allows to reach state of the art VFI on limited motion ranges without relying on additional training data.

Abstract: Recent works have shown the ability of Implicit Neural Representations (INR) to carry meaningful representations of signal derivatives. In this work, we leverage this property to perform Video Frame Interpolation (VFI) by explicitly constraining the derivatives of the INR to satisfy the optical flow constraint equation. We achieve state of the art VFI on limited motion ranges using only a target video and its optical flow, without learning the interpolation operator from additional training data. We further show that constraining the INR derivatives not only allows to better interpolate intermediate frames but also improves the ability of narrow networks to fit the observed frames, which suggests potential applications to video compression and INR optimization.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning