AI Paper Atlas

Research map / Video / Video segmentation and tracking

Segmenting the Motion Components of a Video: A Long-Term Unsupervised Model

Etienne Meunier, Patrick Bouthémy

TPAMI 2025 · Video › Video segmentation and tracking · approach: Learned optical flow estimation · 6 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 the transformer decoder on the 3D U-Net latent space (ablation)LT-MS-K4 · DAVIS2016Jaccard J: 71.3 -> 72.4Improved
add the transformer decoder on the 3D U-Net latent space (ablation)LT-MS-K4 · FBMS59Jaccard J: 55.5 -> 58.2Improved
add the temporal consistency loss term Lc (ablation)LT-MS-K4 · SegTrackV2Jaccard J: 57.5 -> 60.4Improved
add the temporal consistency loss term Lc (ablation)LT-MS-K4 · DAVIS2016Jaccard J: 71.0 -> 72.4Improved
use B-spline temporal motion model instead of polynomial space-time model (ablation)LT-MS-K4 · DAVIS2016Jaccard J: 69.8 -> 72.4Improved
use B-spline temporal motion model instead of polynomial space-time model (ablation)LT-MS-K4 · FBMS59Jaccard J: 54.5 -> 58.2Improved
use the L1-based consistency loss instead of the pre-approximation loss of eq. (11) (ablation)LT-MS-K4 · SegTrackV2Jaccard J: 57.4 -> 60.4Improved
use the L1-based consistency loss instead of the pre-approximation loss of eq. (11) (ablation)LT-MS-K4 · DAVIS2016Jaccard J: 71.6 -> 72.4No clear effect
train on real DAVIS2016 data instead of synthetic FT3DLT-MS-K4 · DAVIS2016, FBMS59, SegTrackV2Jaccard J: 72.4 -> 72.8No clear effect
use long-term LT-MS instead of frame-based EM, sequence-based evaluationLT-MS-K4 · DAVIS2016Jaccard J: 68.0 -> 72.4Improved
use long-term LT-MS instead of frame-based EM, frame-based evaluationLT-MS-K4 · DAVIS2016Jaccard J: 77.1 -> 75.0Hurt
lengthen the input flow sequence at inference from 10LT-MS-K4 · DAVIS2016, FBMS59, SegTrackV2Jaccard JNo clear effect

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