AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

MobileSAM2: Lightweight Segment Anything in Images and Videos via Hypergraphical Knowledge Distillation

Kai Jiang, Jiaxing Huang, Jingyi Zhang, Weiying Xie, Yunsong Li, Yufei Wang, Aoran Xiao, Dacheng Tao

ECCV 2026 · Label-efficient and robust learning › Knowledge distillation · approach: Few-shot and metric distillation

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 HyperKD distillation (ablation)MobileSAM2-23M · LVOS valJ&F: 44.8 -> 69.0Improved
replace FAKD with HyperKDMobileSAM2-23M · LVOS valJ&F: 63.4 -> 69.0Improved
replace FAKD with HyperKDMobileSAM2-5M · LVOS valJ&F: 41.0 -> 48.8Improved
replace FAKD with HyperKDMobileSAM2-10M · LVOS valJ&F: 44.0 -> 51.8Improved
use searched MobileSAM2-23M instead of SAM2-TinyViT-21M with no distillationstudent architecture · LVOS valJ&F: 34.5 -> 44.8Improved
replace FitNet with HyperKDMobileSAM2-23M · LVOS valJ&F: 60.8 -> 69.0Improved
replace two-vertex graph KD (Context Matters, IntRA-KD) with hypergraph KD (ablation)MobileSAM2-23M · LVOS valJ&F: 60.7 -> 69.0Improved
use cosine similarity instead of Euclidean distance for hypergraph distance (ablation)HyperKD · LVOS valJ&F: 67.2 -> 69.0Improved
use cosine similarity instead of Manhattan distance for hypergraph distance (ablation)HyperKD · LVOS valJ&F: 65.7 -> 69.0Improved
set hyperedge threshold epsilon to 1.0 instead of 0.5 (ablation)HyperKD · LVOS valJ&F: 65.2 -> 69.0Improved
set hyperedge threshold epsilon to 1.0 instead of 1.5 (ablation)HyperKD · LVOS valJ&F: 62.0 -> 69.0Improved
vary epsilon from 0.75 to 1.25 (ablation)HyperKD · LVOS valJ&F: 68.4 / 69.0 / 68.1No 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