AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

UL-VIO: Ultra-lightweight Visual-Inertial Odometry with Noise Robust Test-time Adaptation

Jinho Park, Se Young Chun, Mingoo Seok

ECCV 2024 · Label-efficient and robust learning › Test-time adaptation · approach: Test-time adaptation of vision-language models · 7 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
compress NASVIO: pool after last conv in visual encoder, shrink inertial channels, replace LSTM with FCNASVIO · KITTI Seq 05/07/10model size vs relative translation error: 36.45x smaller, t_rel +1.11%, r_rel +1.05 degImproved
replace LSTM with fully connected layers in fused decoder (ablation)UL-VIO · KITTIpose error: 161x smaller D_fused, 1% error increaseNo clear effect
add inertial-pseudo-label BN-only test-time adaptationUL-VIO · KITTI blur noisepose-wise t_rmse [m]: 0.261 -> 0.230Improved
add TTA under multiplicative noiseUL-VIO · KITTI multiplicative noisepose-wise t_rmse [m]: 0.154 -> 0.156No clear effect
use TTA instead of fine-tuning on the matching noiseUL-VIO · KITTI rain noisepose-wise t_rmse [m]: 0.095 (fine-tuned) vs 0.143 (TTA)Hurt
add online TTA with gating under single blur shiftUL-VIO · KITTI Seq 07 blurpose-wise t_rmse [m]: 0.133 -> 0.093Improved
add continual TTA with BN dictionary and domain matchingUL-VIO · KITTI continual shiftspose-wise t_rmse [m]: 0.137 -> 0.113Improved
add continual TTA on EuRoCUL-VIO · EuRoC MH_4_difficultpose-wise t_rmse [m]: 0.0262 -> 0.0254Improved
add continual TTA on real-world Marulan shiftsUL-VIO · Marulanpose-wise t_rmse [m]: 0.237 -> 0.232Improved

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