Multimodal Abstractive Summarization of Instructional Videos with Vision-Language Models

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Nazir, Maham, Aqeel, Muhammad, Zhang, Richong, Setti, Francesco
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911674066272256
author Nazir, Maham
Aqeel, Muhammad
Zhang, Richong
Setti, Francesco
author_facet Nazir, Maham
Aqeel, Muhammad
Zhang, Richong
Setti, Francesco
contents Multimodal video summarization requires visual features that align semantically with language generation. Traditional approaches rely on CNN features trained for object classification, which represent visual concepts as discrete categories not aligned with natural language. We propose ClipSum, a framework that leverages frozen CLIP vision-language features with explicit temporal modeling and dimension-adaptive fusion for instructional video summarization. CLIP's contrastive pre-training on 400M image-text pairs yields visual features semantically aligned with the linguistic concepts that text decoders generate, bridging the vision-language gap at the representation level. On YouCook2, ClipSum achieves 33.0% ROUGE-1 versus 30.5% for ResNet-152 with 4x lower dimensionality (512 vs. 2048), demonstrating that semantic alignment matters more than feature capacity. Frozen CLIP (33.0%) surpasses fine-tuned CLIP (32.3%), showing that preserving pre-trained alignment is more valuable than task-specific adaptation. https://github.com/aqeeelmirza/clipsum
format Preprint
id arxiv_https___arxiv_org_abs_2605_11959
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Multimodal Abstractive Summarization of Instructional Videos with Vision-Language Models
Nazir, Maham
Aqeel, Muhammad
Zhang, Richong
Setti, Francesco
Computer Vision and Pattern Recognition
Computation and Language
Multimodal video summarization requires visual features that align semantically with language generation. Traditional approaches rely on CNN features trained for object classification, which represent visual concepts as discrete categories not aligned with natural language. We propose ClipSum, a framework that leverages frozen CLIP vision-language features with explicit temporal modeling and dimension-adaptive fusion for instructional video summarization. CLIP's contrastive pre-training on 400M image-text pairs yields visual features semantically aligned with the linguistic concepts that text decoders generate, bridging the vision-language gap at the representation level. On YouCook2, ClipSum achieves 33.0% ROUGE-1 versus 30.5% for ResNet-152 with 4x lower dimensionality (512 vs. 2048), demonstrating that semantic alignment matters more than feature capacity. Frozen CLIP (33.0%) surpasses fine-tuned CLIP (32.3%), showing that preserving pre-trained alignment is more valuable than task-specific adaptation. https://github.com/aqeeelmirza/clipsum
title Multimodal Abstractive Summarization of Instructional Videos with Vision-Language Models
topic Computer Vision and Pattern Recognition
Computation and Language
url https://arxiv.org/abs/2605.11959