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

Accurate motion and depth recovery is important for many robot vision tasksincluding autonomous driving. Most previous studies have achieved cooperativemulti-task interaction via either pre-defined loss functions or cross-domainprediction. This paper presents a multi-task scheme that achieves mutualassistance by means of our Flow to Depth (F2D), Depth to Flow (D2F), andExponential Moving Average (EMA). F2D and D2F mechanisms enable multi-scaleinformation integration between optical flow and depth domain based ondifferentiable shallow nets. A dual-head mechanism is used to predict opticalflow for rigid and non-rigid motion based on a divide-and-conquer manner, whichsignificantly improves the optical flow estimation performance. Furthermore, tomake the prediction more robust and stable, EMA is used for our multi-tasktraining. Experimental results on KITTI datasets show that our multi-taskscheme outperforms other multi-task schemes and provide marked improvements onthe prediction results.

A Compacted Structure for Cross-domain learning on Monocular Depth and Flow Estimation