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

Monocular depth inference is a fundamental problem for scene perception ofrobots. Specific robots may be equipped with a camera plus an optional depthsensor of any type and located in various scenes of different scales, whereasrecent advances derived multiple individual sub-tasks. It leads to additionalburdens to fine-tune models for specific robots and thereby high-costcustomization in large-scale industrialization. This paper investigates aunified task of monocular depth inference, which infers high-quality depth mapsfrom all kinds of input raw data from various robots in unseen scenes. A basicbenchmark G2-MonoDepth is developed for this task, which comprises fourcomponents: (a) a unified data representation RGB+X to accommodate RGB plus rawdepth with diverse scene scale/semantics, depth sparsity ([0%, 100%]) anderrors (holes/noises/blurs), (b) a novel unified loss to adapt to diverse depthsparsity/errors of input raw data and diverse scales of output scenes, (c) animproved network to well propagate diverse scene scales from input to output,and (d) a data augmentation pipeline to simulate all types of real artifacts inraw depth maps for training. G2-MonoDepth is applied in three sub-tasksincluding depth estimation, depth completion with different sparsity, and depthenhancement in unseen scenes, and it always outperforms SOTA baselines on bothreal-world data and synthetic data.

G2-MonoDepth: A General Framework of Generalized Depth Inference from Monocular RGB+X Data