AI Paper Atlas

Research map / Video / Video segmentation and tracking

Efficient-SAM2: Accelerating SAM2 with Object-Aware Visual Encoding and Memory Retrieval

Jing Zhang, Zhikai Li, Xuewen Liu, Qingyi Gu

ICLR 2026 · Video › Video segmentation and tracking · approach: Matching-based video object segmentation · 1 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 object-aware Sparse Window Routing to image encoderSAM2.1-B+ · SA-V testJ&F: 77.7 -> 75.0 (1.69x encoder speedup)Hurt
add Sparse Memory Retrieval to memory attentionSAM2.1-B+ · SA-V testJ&F: 77.7 -> 77.8 (1.82x)No clear effect
use SMR saliency mask instead of random mask (ablation)SMR · SA-V test (B+)J&F: 76.7 -> 77.8Improved
use SMR saliency mask instead of uniform mask (ablation)SMR · SA-V test (B+)J&F: 76.8 -> 77.8Improved
add learned shortcut branch instead of identity bypass (ablation)SWR · SA-V (SAM2.1-B+)J&F: 73.0 -> 75.0Improved
use bottleneck FFN instead of full FFN shortcut (ablation)SWR · SA-V (SAM2.1-L)J&F: 79.2 -> 79.0No clear effect
add mask dilation to window selection (ablation)SWR · SA-V test (B+)J&F: 73.6 -> 75.0Improved
include all decoder predictions in window selection (ablation)SWR · SA-V test (B+)J&F: 73.6 -> 74.1Improved
raise memory sparsity beyond about 0.95 (ablation)SMR · VOS accuracyJ&FHurt
combine SWR with SMREfficient-SAM2 · SA-V test (SAM2.1-B+, dt=1)J&F: 77.7 -> 75.5 (1.74x)Hurt
use token merging (ToMe) for image encoder instead of SWRSAM2.1-B+ · SA-V testJ&F: 77.7 -> 55.3 (1.36x)Hurt

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