AI Paper Atlas

Research map / Video / Video-language understanding

Bridging VideoQA and Video-Guided Agentic Tasks via Generalized Keyframe Extraction

Sunqi Fan, Qingle Liu, Runqi Yin, Meng-Hao Guo, Shuojin Yang

ECCV 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 0 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
replace BFS expansion with A* cost-function search (ablation)TASKER (GPT-4) · EgoSchema subsetaccuracy: 64.7 -> 68.0Improved
combine task-driven and scene-aware costs (A*) versus either alone (ablation)TASKER · EgoSchema subsetaccuracy: 67.0 / 66.8 -> 68.0Improved
use A* instead of other variants on VG-GUI-BenchTASKER · VG-GUI-Benchaccuracy: 40.01 -> 40.96Improved
A* variant on VG-GUI-Bench completion and type accuracy (ablation)TASKER-A* · VG-GUI-Benchcompletion: 71.38 vs 74.39 (Dijkstra)Mixed
combine self-evaluation and summarization confidence by voting (ablation)TASKER-A* · EgoSchema subsetaccuracy: 67.4 / 67.3 -> 68.0Improved
use GPT-4o as base LLM instead of GPT-4 (ablation)TASKER · EgoSchema subsetaccuracy: 68.0 -> 68.6Improved
use reasoning models (o3-mini, Deepseek-R1) as base LLM (ablation)TASKER · EgoSchema subsetaccuracy: 68.0 -> 67.3 / 67.6Hurt
use TASKER keyframe search instead of VideoTreeGPT-4 backbone · NExT-QA averageaccuracy: 75.6 -> 77.4Improved
use TASKER instead of VideoTree on Qwen3-VLQwen3-VL · EgoSchema fullsetaccuracy: 76.7 -> 77.3Improved
add 10 uniformly sampled video frames versus no videofrontier MLLMs · VG-GUI-Benchaccuracy: 35.93 -> 39.78 (Seed-2.0-Pro)Improved
feed oracle keyframes instead of selected framesQwen3-VL-235B-A22B · VG-GUI-Benchaccuracy: 44.32 vs 40.96Mixed
add TASKER-selected frames versus uniform frames on OSWorld subsetGemini-3-Flash · OSWorld subsettask success rate: 49.13 -> 50.63Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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