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

Single-view depth estimation from omnidirectional images has gainedpopularity with its wide range of applications such as autonomous driving andscene reconstruction. Although data-driven learning-based methods demonstratesignificant potential in this field, scarce training data and ineffective 360estimation algorithms are still two key limitations hindering accurateestimation across diverse domains. In this work, we first establish alarge-scale dataset with varied settings called Depth360 to tackle the trainingdata problem. This is achieved by exploring the use of a plenteous source ofdata, 360 videos from the internet, using a test-time training method thatleverages unique information in each omnidirectional sequence. With novelgeometric and temporal constraints, our method generates consistent andconvincing depth samples to facilitate single-view estimation. We then proposean end-to-end two-branch multi-task learning network, SegFuse, that mimics thehuman eye to effectively learn from the dataset and estimate high-quality depthmaps from diverse monocular RGB images. With a peripheral branch that usesequirectangular projection for depth estimation and a foveal branch that usescubemap projection for semantic segmentation, our method predicts consistentglobal depth while maintaining sharp details at local regions. Experimentalresults show favorable performance against the state-of-the-art methods.

360 Depth Estimation in the Wild -- The Depth360 Dataset and the SegFuse Network