SOURCE-LINKED INTELLIGENCE

Evaluation of MLLM-Agnostic Plug-and-Play Keyframe Selection Methods for Long Video Understanding

arXiv · AI, language, vision and robotics · article · Sep 4, 2026 · UTC

Multimodal large language models (MLLMs) cannot process every frame of a long video because of limitations in visual-token and computational budgets. Three primary approaches have been proposed to enhance their long-video understanding capabilities: (i) Retraining an MLLM on a large video corpus and/or extending its input length; (ii) Training an adapter for a specific MLLM that takes the entire video and the query as input and selects the most relevant video frames; and (iii) Developing a training-free, plug-and-play (PaP) adapter that is MLLM-agnostic. We refer to the third approach as PaP k

Read original source ↗ Open in workspace

recordType

paper

region

Global