AI Paper Atlas

Research map / Video / Video-language understanding

TimeLens: Rethinking Video Temporal Grounding with Multimodal LLMs

Jun Zhang, Teng Wang, Yuying Ge, Yixiao Ge, Xinhao Li, Limin Wang

CVPR 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 38 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
re-annotate benchmarks with strict criteriaVTG models · Charades-STAMean IoU: re-rankingMixed
use interleaved textual timestamp prefix with raw timestamps instead of visual overlay (ablation)Qwen2.5-VL-7B · ActivityNet-TimeLensmIoU: 39.8 -> 43.1Improved
use interleaved rather than non-interleaved textual prefix (ablation)Qwen2.5-VL-7B · QVHighlights-TimeLensmIoU: 42.8 -> 56.7Improved
use raw timestamps instead of frame indices (ablation)Qwen2.5-VL-7B · ActivityNet-TimeLensmIoU: 36.9 -> 43.1Improved
replace position-embedding time encoding with interleaved text (ablation)Qwen2.5-VL-7B · Charades-TimeLensmIoU: 36.6 -> 48.3Improved
thinking-free RLVR instead of SFT on 100K data (ablation)Qwen2.5-VL-7B · ActivityNet-TimeLensmIoU: 39.7 -> 43.1Improved
remove thinking from RLVR (ablation)Qwen2.5-VL-7B · Charades-TimeLensmIoU: 42.7 -> 48.3Improved
add SFT phase before RLVR (ablation)Qwen2.5-VL-7B · Charades-TimeLensmIoU: 48.3 -> 50.1No clear effect
continue RLVR training past reward plateau (ablation)Qwen2.5-VL-7B · TimeLens-BenchperformanceHurt
sample higher-difficulty training data for RLVR (ablation)Qwen2.5-VL-7B · TimeLens-BenchperformanceImproved
train on TimeLens-100K re-annotated dataQwen2.5-VL-7B · TimeLens-BenchperformanceImproved

Later papers that compared with this method

2 later accepted papers report TimeLens-8B as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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