AI Paper Atlas

Research map / Video / Video segmentation and tracking

CaptionFormer: Unified Segmentation, Tracking, and Captioning for Spatio-Temporal Objects

Gabriel Fiastre, Antoine Yang, Cordelia Schmid

CVPR 2026 · Video › Video segmentation and tracking · approach: Matching-based video object segmentation · 3 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 LVISCap image pretraining to LV-VISCap training (ablation)CaptionFormer · LV-VISCap valCapA: 37.9 -> 43.6Improved
add LVISCap and LV-VISCap pretraining (vs COCO only)CaptionFormer · VidSTG DVOC valCapA: 44.3 -> 51.0Improved
add query temporal aggregation (32 clips) (ablation)CaptionFormer · VidSTG DVOC valCapA: 51.0 -> 55.4Improved
add temporal aggregation (ablation)CaptionFormer · VLN DVOC valCapA: 22.9 -> 23.4No clear effect
add temporal aggregation (ablation)CaptionFormer · BenSMOT valCIDEr: 39.9 -> 42.6Improved
add mask loss in detection loss (ablation)CaptionFormer · VLN DVOC valCHOTA: 45.4 -> 47.4Improved
weighted mean instead of arithmetic mean for aggregation (ablation)CaptionFormer · VidSTG DVOC valCapA: 49.1 -> 51.6Improved
add captioning loss (ablation)CaptionFormer · LV-VISCap valVIS mAP: 31.7 -> 34.2Improved
add bounding-box visual prompt plus category labels in text prompt (ablation)CaptionFormer · caption generation (Gemini)avg rating: 29.5 -> 80.7Improved
use mask boundaries instead of bounding boxes as visual prompt (ablation)CaptionFormer · caption generation (Gemini)avg rating: 85.1 -> 77.1Hurt
add LVISCap captions (ablation)CaptionFormer · LV-VISCap manual subsetCapA: 22.9 -> 31.3Improved
use best-score single clip vs middle frame (ablation)CaptionFormer · VidSTG DVOC valCapA: 46.9 -> 51.0Improved

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