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

Self-supervised deep learning methods have leveraged stereo images fortraining monocular depth estimation. Although these methods show strong resultson outdoor datasets such as KITTI, they do not match performance of supervisedmethods on indoor environments with camera rotation. Indoor, rotated scenes arecommon for less constrained applications and pose problems for two reasons:abundance of low texture regions and increased complexity of depth cues forimages under rotation. In an effort to extend self-supervised learning to moregeneralised environments we propose two additions. First, we propose a novelFilled Disparity Loss term that corrects for ambiguity of image reconstructionerror loss in textureless regions. Specifically, we interpolate disparity inuntextured regions, using the estimated disparity from surrounding texturedareas, and use L1 loss to correct the original estimation. Our experiments showthat depth estimation is substantially improved on low-texture scenes, withoutany loss on textured scenes, when compared to Monodepth by Godard et al.Secondly, we show that training with an application's representative rotations,in both pitch and roll, is sufficient to significantly improve performance overthe entire range of expected rotation. We demonstrate that depth estimation issuccessfully generalised as performance is not lost when evaluated on test setswith no camera rotation. Together these developments enable a broader use ofself-supervised learning of monocular depth estimation for complexenvironments.

Self-Supervised Monocular Depth Estimation of Untextured Indoor Rotated Scenes