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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Self-supervised monocular depth prediction provides a cost-effective solutionto obtain the 3D location of each pixel. However, the existing approachesusually lead to unsatisfactory accuracy, which is critical for autonomousrobots. In this paper, we propose FusionDepth, a novel two-stage network toadvance the self-supervised monocular dense depth learning by leveraginglow-cost sparse (e.g. 4-beam) LiDAR. Unlike the existing methods that usesparse LiDAR mainly in a manner of time-consuming iterative post-processing,our model fuses monocular image features and sparse LiDAR features to predictinitial depth maps. Then, an efficient feed-forward refine network is furtherdesigned to correct the errors in these initial depth maps in pseudo-3D spacewith real-time performance. Extensive experiments show that our proposed modelsignificantly outperforms all the state-of-the-art self-supervised methods, aswell as the sparse-LiDAR-based methods on both self-supervised monocular depthprediction and completion tasks. With the accurate dense depth prediction, ourmodel outperforms the state-of-the-art sparse-LiDAR-based method(Pseudo-LiDAR++) by more than 68% for the downstream task monocular 3D objectdetection on the KITTI Leaderboard. Code is available athttps://github.com/AutoAILab/FusionDepth

Advancing Self-supervised Monocular Depth Learning with Sparse LiDAR