AI Paper Atlas

Research map / Video / Video-language understanding

HumanSAM: Classifying Human-centric Forgery Videos in Human Spatial, Appearance, and Motion Anomaly

Chang Liu, Yunfan Ye, Fan Zhang, Qingyang Zhou, Yuchuan Luo, Zhiping Cai

ICCV 2025 · Video › Video-language understanding · approach: Event and anomaly reasoning in video · 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
use HumanSAM dual-branch model instead of TimeSformer for 4-class forgery classificationHumanSAM · HFV multi-classACC (%): 64.0 -> 69.8Improved
add the Cat module combining attention-pooled and average-pooled features (ablation)InternVideo2 backbone · HFV multi-classACC (%): 63.3 -> 65.6 ACC; 87.4 -> 86.3 AUCMixed
add monocular depth branch (Depth Pro) (ablation)InternVideo2 + Cat · HFV multi-classACC (%): 65.6 -> 69.2Improved
add rank-based confidence loss (ablation)InternVideo2 + Cat + Depth · HFV multi-classACC (%): 69.2 -> 69.8Improved
train on multi-class labels instead of binary on weak detectors (Uni-FD)Uni-FD · HFV binaryACC (%): 80.9 -> 92.3Improved
train on multi-class labels instead of binary on strong detectorsTimeSformer, MM-Det · HFV binaryACC (%): 95.2 -> 94.8 (TimeSformer)No clear effect
apply Gaussian blur sigma=3 to test videosHumanSAM · CogVideoX-5B multi-classAUC (%): 92.1 -> 88.4Hurt
apply JPEG or resize post-processing to test videosHumanSAM · CogVideoX-5B multi-classAUC (%): 92.1 -> 91.8 (JPEG), 91.9 (resize)No clear effect
train on MiniMax or Kling forgeries instead of lower-ranked CogVideoX-2B (ablation)HumanSAM · HFV multi-classACC (%): 69.8 -> 66.5 (MinMax), 64.1 (Kling)Hurt
HumanSAM versus MM-Det on Sora external datasetHumanSAM · SoraACC (%): 81.0 -> 95.3Improved

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