PhaseNet for Video Frame Interpolation

TL;DR

PhaseNet introduces a neural network that estimates phase decomposition for video frame interpolation, effectively handling challenging scenarios with lighting changes, motion blur, and larger motions, outperforming previous methods.

Contribution

It presents a novel neural network decoder for phase-based motion representation, improving robustness and motion handling in video frame interpolation.

Findings

Outperforms recent deep learning approaches on challenging datasets.

Robustly handles lighting changes and motion blur.

Effectively manages larger motions than previous phase-based methods.

Abstract

Most approaches for video frame interpolation require accurate dense correspondences to synthesize an in-between frame. Therefore, they do not perform well in challenging scenarios with e.g. lighting changes or motion blur. Recent deep learning approaches that rely on kernels to represent motion can only alleviate these problems to some extent. In those cases, methods that use a per-pixel phase-based motion representation have been shown to work well. However, they are only applicable for a limited amount of motion. We propose a new approach, PhaseNet, that is designed to robustly handle challenging scenarios while also coping with larger motion. Our approach consists of a neural network decoder that directly estimates the phase decomposition of the intermediate frame. We show that this is superior to the hand-crafted heuristics previously used in phase-based methods and also compares favorably to recent deep learning based approaches for video frame interpolation on challenging datasets.