High Resolution Multi-Scale RAFT (Robust Vision Challenge 2022)
This work addresses optical flow estimation for computer vision applications, representing an incremental improvement over existing methods.
The authors tackled optical flow estimation by extending the MS-RAFT method with an additional finer scale and on-demand cost computation, achieving first place overall in the Robust Vision Challenge 2022, with top rankings on benchmarks like VIPER and KITTI.
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.