AI Paper Atlas

Research map / 3D vision / Depth and stereo

"JPerceiver: Joint Perception Network for Depth, Pose and Layout Estimation in Driving Scenes"

Haimei Zhao, Jing Zhang, Sen Zhang, Dacheng Tao

ECCV 2022 · 3D vision › Depth and stereo · approach: Self-supervised monocular depth · 19 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add cross-view transfer (CCT-CV) (ablation)baseline 1-1 · Argoverse road layoutmIoU: 76.66 -> 77.76Improved
add cross-modal transfer (CCT-CM) on top of CCT-CV (ablation)baseline 1-1 · Argoverse road layoutmAP: 88.42 -> 89.00Improved
add complete CCT (ablation)baseline 1-1 · Argoverse vehicle layoutmIoU: 46.97 -> 49.39Improved
add hybrid loss (ablation)1-1 with CCT · Argoverse road layoutmAP: 89.00 -> 90.59Improved
share one encoder for road and vehicle layouts (ablation)baseline · Argoverse vehicle layoutmIoU: 46.97 -> 42.77Hurt
add CCT and hybrid loss to shared-encoder baseline (ablation)baseline 1-2 · Argoverse vehicle layoutmIoU: 42.77 -> 49.45Improved
use shared encoder for depth and layout networks (ablation)JPerceiver · KITTI Odometry road layoutmIoU: 78.13 -> 77.53Hurt
use shared encoder for depth and layout networks (ablation)JPerceiver · KITTI Odometry VO sequence 07t_err (%): 4.57 -> 9.73Hurt
use shared encoder for depth and layout networks (ablation)JPerceiver · KITTI Odometry depth (1024x1024)Abs Rel: 0.116 -> 0.133Hurt
add CCT (ablation)depth baseline · KITTI Odometry depth (512x512)Abs Rel: 0.120 -> 0.108Improved
add CGT scale loss (ablation)depth baseline · KITTI Odometry depth (512x512)Abs Rel: 0.120 -> 0.135Hurt
add CGT scale loss (ablation)depth baseline · KITTI Odometry depth (512x512)scale factor: 39.452 -> 1.088Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy