AI Paper Atlas

Research map / 3D vision / Depth and stereo

SemARFlow: Injecting Semantics into Unsupervised Optical Flow Estimation for Autonomous Driving

Shuai Yuan, Shuzhi Yu, Hannah Kim, Carlo Tomasi

ICCV 2023 · 3D vision › Depth and stereo · approach: Learned optical and scene flow · 10 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 semantic segmentation inputs, encoder and augmentation, to adapted ARFlowARFlow · KITTI-2015 testFl-all (%): 11.80 -> 8.38Improved
add semantic encoder to baseline (ablation)SemARFlow baseline · KITTI-2015 testFl-all (%): 9.89 -> 8.99Improved
add semantic augmentation on top of semantic encoder (ablation)SemARFlow · KITTI-2015 testFl-all (%): 8.99 -> 8.38Improved
add learned upsampler while keeping smoothness loss (ablation)ARFlow · KITTI-2015 trainFl-all (%): 10.36 -> 10.22No clear effect
turn off smoothness loss together with learned upsampler (ablation)ARFlow with upsampler · KITTI-2015 trainFl-all (%): 10.22 -> 8.87Improved
add semantic encoder at level 3 versus other levels (ablation)ARFlow · KITTI-2015 trainFl-all (%): 10.36 -> 9.75Improved
add semantic encoder on upsampler/no-smoothness model (ablation)SemARFlow · KITTI-2015 trainFl-all (%): 8.87 -> 8.26Improved
add semantic augmentation on upsampler/no-smoothness model (ablation)SemARFlow · KITTI-2015 trainFl-all (%): 8.87 -> 8.80Improved
start semantic augmentation at 150k rather than 100k iterations (ablation)SemARFlow · KITTI-2015 trainFl-all (%): 7.92 -> 7.79Improved
augment poles and sky in addition to vehicles (ablation)SemARFlow · KITTI-2015 trainFl-all (%): 7.94 -> 7.79Improved
apply augmentation loss on full region rather than only newly occluded region (ablation)SemARFlow · KITTI-2015 trainFl-all (%): 8.15 -> 7.79Improved
add semantic inputs for cross-dataset trainingSemARFlow · Cityscapes to KITTI-2015 trainFl-all (%): 12.27 -> 10.32Improved

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