Frame Interpolation with Multi-Scale Deep Loss Functions and Generative Adversarial Networks

TL;DR

This paper introduces FIGAN, a multi-scale GAN for efficient and high-quality video frame interpolation, combining a novel residual estimation module with perceptual loss functions to outperform existing methods.

Contribution

The paper proposes a novel multi-scale residual estimation module and a combined adversarial and content loss for improved frame interpolation quality and efficiency.

Findings

Achieves state-of-the-art accuracy in frame interpolation

Provides subjective visual quality comparable to top methods

Runs 47 times faster than existing approaches

Abstract

Frame interpolation attempts to synthesise frames given one or more consecutive video frames. In recent years, deep learning approaches, and notably convolutional neural networks, have succeeded at tackling low- and high-level computer vision problems including frame interpolation. These techniques often tackle two problems, namely algorithm efficiency and reconstruction quality. In this paper, we present a multi-scale generative adversarial network for frame interpolation (\mbox{FIGAN}). To maximise the efficiency of our network, we propose a novel multi-scale residual estimation module where the predicted flow and synthesised frame are constructed in a coarse-to-fine fashion. To improve the quality of synthesised intermediate video frames, our network is jointly supervised at different levels with a perceptual loss function that consists of an adversarial and two content losses. We evaluate the proposed approach using a collection of 60fps videos from YouTube-8m. Our results improve the state-of-the-art accuracy and provide subjective visual quality comparable to the best performing interpolation method at x47 faster runtime.