English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Optical flow estimation is a challenging problem remaining unsolved. Recentdeep learning based optical flow models have achieved considerable success.However, these models often train networks from the scratch on standard opticalflow data, which restricts their ability to robustly and geometrically matchimage features. In this paper, we propose a rethinking to previous optical flowestimation. We particularly leverage Geometric Image Matching (GIM) as apre-training task for the optical flow estimation (MatchFlow) with betterfeature representations, as GIM shares some common challenges as optical flowestimation, and with massive labeled real-world data. Thus, matching staticscenes helps to learn more fundamental feature correlations of objects andscenes with consistent displacements. Specifically, the proposed MatchFlowmodel employs a QuadTree attention-based network pre-trained on MegaDepth toextract coarse features for further flow regression. Extensive experiments showthat our model has great cross-dataset generalization. Our method achieves11.5% and 10.1% error reduction from GMA on Sintel clean pass and KITTI testset. At the time of anonymous submission, our MatchFlow(G) enjoysstate-of-the-art performance on Sintel clean and final pass compared topublished approaches with comparable computation and memory footprint. Codesand models will be released in https://github.com/DQiaole/MatchFlow.

Rethinking Optical Flow from Geometric Matching Consistent Perspective