AI Paper Atlas

Research map / 3D vision / 3D generation

Veritatem Dies Aperit - Temporally Consistent Depth Prediction Enabled by a Multi-Task Geometric and Semantic Scene Understanding Approach

Amir Atapour-Abarghouei, Toby P. Breckon

CVPR 2019 · 3D vision › 3D generation · approach: Semantic scene completion · 42 citations (OpenAlex)

Check 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
jointly train depth prediction and segmentation in one network instead of separate networks (ablation)multi-task scene understanding model · Random synthetic sequencesAbs. Rel.: 0.245 -> 0.208Improved
jointly train depth prediction and segmentation in one network instead of separate networks (ablation)multi-task scene understanding model · Random synthetic sequencesIoU: 0.672 -> 0.764Improved
add optical flow supervision (ablation)T/R/A/SC/S depth model · Random synthetic sequencesRMSE: 6.256 -> 6.026Improved
add smoothing loss and optical flow supervision (ablation)T/R/A/SC depth model · Random synthetic sequencesRMSE: 6.473 -> 6.026Improved
add skip connections, smoothing loss, and optical flow supervision (ablation)T/R/A depth model · Random synthetic sequencesRMSE: 6.826 -> 6.026Improved
add adversarial loss, skip connections, smoothing loss, and optical flow supervision (ablation)T/R depth model · Random synthetic sequencesRMSE: 7.393 -> 6.026Improved
use temporal training and optical flow supervision instead of non-temporal training without optical flow (ablation)depth and segmentation model · Random synthetic sequencesRMSE: 6.469 -> 6.026Improved
use the joint model rather than separate models (ablation)scene depth and segmentation · Random synthetic sequencesdepth accuracy (sigma < 1.25): 0.803 -> 0.836Improved
add the proposed method rather than DPNCityscapes segmentation · Cityscapes testIoU: 66.8 -> 67.0Improved
use the proposed method rather than ResegCamVid segmentation · CamVid testIoU: 58.8 -> 59.1Improved
use the proposed method rather than FDFdepth completion · Synthetic depth completion test sequencesPSNR: 46.13 -> 47.45Improved

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