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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Optical flow estimation is crucial for various applications in vision androbotics. As the difficulty of collecting ground truth optical flow inreal-world scenarios, most of the existing methods of learning optical flowstill adopt synthetic dataset for supervised training or utilize photometricconsistency across temporally adjacent video frames to drive the unsupervisedlearning, where the former typically has issues of generalizability while thelatter usually performs worse than the supervised ones. To tackle suchchallenges, we propose to leverage the geometric connection between opticalflow estimation and stereo matching (based on the similarity upon finding pixelcorrespondences across images) to unify various real-world depth estimationdatasets for generating supervised training data upon optical flow.Specifically, we turn the monocular depth datasets into stereo ones viasynthesizing virtual disparity, thus leading to the flows along the horizontaldirection; moreover, we introduce virtual camera motion into stereo data toproduce additional flows along the vertical direction. Furthermore, we proposeapplying geometric augmentations on one image of an optical flow pair,encouraging the optical flow estimator to learn from more challenging cases.Lastly, as the optical flow maps under different geometric augmentationsactually exhibit distinct characteristics, an auxiliary classifier which trainsto identify the type of augmentation from the appearance of the flow map isutilized to further enhance the learning of the optical flow estimator. Ourproposed method is general and is not tied to any particular flow estimator,where extensive experiments based on various datasets and optical flowestimation models verify its efficacy and superiority.

Skin the sheep not only once: Reusing Various Depth Datasets to Drive the Learning of Optical Flow