AI Paper Atlas

Research map / Video / Video-language understanding

X2SAM: Any Segmentation in Images and Videos

Hao Wang, Limeng Qiao, Chi Zhang, Lin Ma, Guanglu Wan, Xiangyuan Lan, Xiaodan Liang

ECCV 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 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 zero-initialized token-to-image attention to mask decoder (ablation)X2SAM decoder · V-Ref. YT21J&F: 53.6 -> 60.8Improved
add randomly initialized token-to-image attention (ablation)X2SAM decoder · I-Ref. RefCOCO/+/gcIoU: 82.9/78.0/79.5 -> 82.5/77.2/79.4Hurt
use zero init instead of random init for T2I attention (ablation)X2SAM decoder · V-Ref. YT21J&F: 58.2 -> 60.8Improved
add single-scale mask memory (ablation)X2SAM · V-Ref. YT21J&F: 53.6 -> 52.5No clear effect
add mask guidance to memory (ablation)X2SAM memory · V-Ref. YT21J&F: 53.6 -> 63.3Improved
add multi-scale memory design (ablation)X2SAM memory · V-Rea. ReVOS AllJ&F: 36.5 -> 53.5Improved
increase memory size from 1 to 6 (ablation)X2SAM · V-OV YT-VIS21mAP: 58.5 -> 60.2Improved
increase memory size to 8 (ablation)X2SAM · V-Rea. ReVOS AllJ&F: 57.5 -> 53.5Hurt
use unified joint training instead of simple joint training (ablation)X2SAM · training costGPU hours: 5.2K -> 3.3KImproved
use unified joint training instead of separate training (ablation)X2SAM · V-Gen. mIoUmIoU: 61.1 -> 64.7Improved
extend image model to video with mask memory (X2SAM vs X-SAM)X2SAM · I-Gen./I-VGD image tasksPQ/AP: I-OV 20.9 -> 31.2 PQ; I-Gen. 54.7 -> 54.1 PQ; I-VGD COCO point 47.9 -> 45.9Mixed
X2SAM vs video MLLM generalist on grounded conversationX2SAM · V-GCGmIoU: 54.3 -> 75.8Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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