AI Paper Atlas

Research map / Video / Video segmentation and tracking

VIL-100: A New Dataset and a Baseline Model for Video Instance Lane Detection

Yujun Zhang, Lei Zhu, Wei Feng, Huazhu Fu, Mingqian Wang, Qingxia Li, Cheng Li, Song Wang

ICCV 2021 · Video › Video segmentation and tracking · approach: Self-supervised video representation learning · 73 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 attention to the local memory (ablation)Basic (STM-like) · VIL-100 testmIoU: 0.638 -> 0.688Improved
add attentive global memory (ablation)Basic (STM-like) · VIL-100 testmIoU: 0.638 -> 0.670Improved
add global memory to local memory (ablation)+LM · VIL-100 testmIoU: 0.688 -> 0.693Improved
add local memory to global memory (ablation)+GM · VIL-100 testmIoU: 0.670 -> 0.693Improved
add multi-level LGMA modules on low-level and high-level features (ablation)+LGM · VIL-100 testmIoU: 0.693 -> 0.705Improved
add multi-level LGMA modules (ablation)+LGM · VIL-100 testM_T (video-based): 0.760 -> 0.764Improved
add local and global memory attention (full model) to Basic (ablation)Basic (STM-like) · VIL-100 testF1 (IoU>0.8): 0.402 -> 0.458Improved
use 5 sampled frames instead of 3 (ablation)MMA-Net · VIL-100 testmIoU: 0.678 -> 0.715Improved
use 7 sampled frames instead of 5 (ablation)MMA-Net · VIL-100 testmIoU: 0.715 -> 0.705No clear effect
MMA-Net versus best baseline STM on video-based M_TMMA-Net · VIL-100 testM_T: 0.656 -> 0.764Improved
MMA-Net versus best baseline on F1 (IoU>0.8)MMA-Net · VIL-100 testF1 (IoU>0.8): 0.327 -> 0.458Improved

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