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/

Zendy Plus

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

Zendy Tools

Zendy Open

Online stereo adaptation tackles the domain shift problem, caused bydifferent environments between synthetic (training) and real (test) datasets,to promptly adapt stereo models in dynamic real-world applications such asautonomous driving. However, previous methods often fail to counteractparticular regions related to dynamic objects with more severe environmentalchanges. To mitigate this issue, we propose to incorporate an auxiliarypoint-selective network into a meta-learning framework, called PointFix, toprovide a robust initialization of stereo models for online stereo adaptation.In a nutshell, our auxiliary network learns to fix local variants intensivelyby effectively back-propagating local information through the meta-gradient forthe robust initialization of the baseline model. This network ismodel-agnostic, so can be used in any kind of architectures in a plug-and-playmanner. We conduct extensive experiments to verify the effectiveness of ourmethod under three adaptation settings such as short-, mid-, and long-termsequences. Experimental results show that the proper initialization of the basestereo model by the auxiliary network enables our learning paradigm to achievestate-of-the-art performance at inference.

PointFix: Learning to Fix Domain Bias for Robust Online Stereo Adaptation