High Resolution Multi-Scale RAFT (Robust Vision Challenge 2022)

TL;DR

This paper introduces MS-RAFT+, a multi-scale optical flow method that achieved top rankings in the Robust Vision Challenge 2022 by combining multi-scale concepts with on-demand cost computation and improved training for better generalization.

Contribution

The paper presents MS-RAFT+, an extension of MS-RAFT, integrating an additional finer scale and a shared convex upsampler, enhancing accuracy and generalization in optical flow estimation.

Findings

Ranked first on VIPER benchmark

Second on KITTI, Sintel, Middlebury

Achieved overall first place in the challenge

Abstract

In this report, we present our optical flow approach, MS-RAFT+, that won the Robust Vision Challenge 2022. It is based on the MS-RAFT method, which successfully integrates several multi-scale concepts into single-scale RAFT. Our approach extends this method by exploiting an additional finer scale for estimating the flow, which is made feasible by on-demand cost computation. This way, it can not only operate at half the original resolution, but also use MS-RAFT's shared convex upsampler to obtain full resolution flow. Moreover, our approach relies on an adjusted fine-tuning scheme during training. This in turn aims at improving the generalization across benchmarks. Among all participating methods in the Robust Vision Challenge, our approach ranks first on VIPER and second on KITTI, Sintel, and Middlebury, resulting in the first place of the overall ranking.