AI Paper Atlas

Research map / Video / Video segmentation and tracking

MOSE: A New Dataset for Video Object Segmentation in Complex Scenes

Henghui Ding, Chang Liu, Shuting He, Xudong Jiang, Philip H.S. Torr, Song Bai

ICCV 2023 · Video › Video segmentation and tracking · approach: Matching-based video object segmentation · 176 citations (OpenAlex)

Read the paper

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
evaluate state-of-the-art semi-supervised VOS on complex scenes (MOSE) instead of DAVIS 2017DeAOT · MOSE val vs DAVIS 2017J&F: 85.2 -> 59.4Hurt
evaluate XMem on MOSE instead of DAVIS 2017XMem · MOSE val vs DAVIS 2017J&F: 86.2 -> 57.6Hurt
use bounding-box instead of mask first-frame initializationsemi-supervised VOS · MOSE valJ&F: box methods 22.0 and 23.8 vs mask methods 48.8 to 59.4Hurt
evaluate interactive VOS on MOSE instead of DAVIS 2017STCN · MOSE val vs DAVIS 2017J&F@60s: 88.8 -> 56.8Hurt
evaluate zero-shot multi-object VOS on MOSE instead of DAVIS 2017STEm-Seg · MOSE val vs DAVIS 2017J&F: 64.7 -> 46.9Hurt
use mBOR as occlusion indicatorMOSE dataset analysis · occlusion measurementindicator reliabilityHurt

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