AI Paper Atlas

Research map / Video / Video-language understanding

PerceptionComp: A Video Benchmark for Complex Perception-Centric Reasoning

Shaoxuan Li, Zhixuan Zhao, Hanze Deng, Zirun Ma, Shulin Tian, Zuyan Liu, Yushi Hu, Haoning Wu, Yuhao Dong, Benlin Liu, Ziwei Liu, Ranjay Krishna

ECCV 2026 · Video › Video-language understanding · approach: Event and anomaly reasoning in video · 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
use thinking model GPT-o3 instead of GPT-4oGPT-4o · PerceptionCompaccuracy: +11.04%Improved
use Gemini-2.5-Pro instead of Gemini-2.5-FlashGemini-2.5-Flash · PerceptionCompaccuracy: +6.19%Improved
use thinking variant of Qwen3-VL-8BQwen3-VL-8B · PerceptionCompaccuracy: 34.06 -> 33.82No clear effect
use reasoning-trained VideoChat-R1 7B instead of Qwen2.5-VL 7BQwen2.5-VL-7B · PerceptionCompaccuracy: 22.73 -> 28.63Improved
use reasoning-trained Video-R1 7B instead of Qwen2.5-VL 7BQwen2.5-VL-7B · PerceptionCompaccuracy: 22.73 -> 26.27Improved
increase input frames from 16 to 50 (ablation)GPT-o3 · PerceptionComp (100-video subset)accuracy: 34.0 -> 43.54Improved
increase input frames from 16 to 64 (ablation)Qwen3-VL-8B · PerceptionComp (100-video subset)accuracy: 27.0 -> 34.80Improved
increase thinking-token budget from 1,024 to 8,192 (ablation)Gemini-2.5-Flash · PerceptionComp (100-video subset)accuracyImproved
scale Qwen3-VL from 8B to 235BQwen3-VL · PerceptionCompaccuracy: 34.06 -> 34.02No clear effect
scale Qwen2.5-VL from 7B to 72BQwen2.5-VL · PerceptionCompaccuracy: 22.73 -> 31.33Improved
scale Qwen3-VL-Thinking from 8B to 235BQwen3-VL-Thinking · PerceptionCompaccuracy: 33.82 -> 38.20Improved
allow unrestricted rewatching for humans (ablation)Human · PerceptionCompaccuracy: 18.97 -> 85.10Improved

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