Saved in:
Bibliographic Details
Main Authors: Tzachor, Issar, Samuel, Dvir, Ben-Ari, Rami
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.08099
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917257992470528
author Tzachor, Issar
Samuel, Dvir
Ben-Ari, Rami
author_facet Tzachor, Issar
Samuel, Dvir
Ben-Ari, Rami
contents Recent studies have adapted generative Multimodal Large Language Models (MLLMs) into embedding extractors for vision tasks, typically through fine-tuning to produce universal representations. However, their performance on video remains inferior to Video Foundation Models (VFMs). In this paper, we focus on leveraging MLLMs for video-text embedding and retrieval. We first conduct a systematic layer-wise analysis, showing that intermediate (pre-trained) MLLM layers already encode substantial task-relevant information. Leveraging this insight, we demonstrate that combining intermediate-layer embeddings with a calibrated MLLM head yields strong zero-shot retrieval performance without any training. Building on these findings, we introduce a lightweight text-based alignment strategy which maps dense video captions to short summaries and enables task-related video-text embedding learning without visual supervision. Remarkably, without any fine-tuning beyond text, our method outperforms current methods, often by a substantial margin, achieving state-of-the-art results across common video retrieval benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2602_08099
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle VidVec: Unlocking Video MLLM Embeddings for Video-Text Retrieval
Tzachor, Issar
Samuel, Dvir
Ben-Ari, Rami
Computer Vision and Pattern Recognition
Artificial Intelligence
Recent studies have adapted generative Multimodal Large Language Models (MLLMs) into embedding extractors for vision tasks, typically through fine-tuning to produce universal representations. However, their performance on video remains inferior to Video Foundation Models (VFMs). In this paper, we focus on leveraging MLLMs for video-text embedding and retrieval. We first conduct a systematic layer-wise analysis, showing that intermediate (pre-trained) MLLM layers already encode substantial task-relevant information. Leveraging this insight, we demonstrate that combining intermediate-layer embeddings with a calibrated MLLM head yields strong zero-shot retrieval performance without any training. Building on these findings, we introduce a lightweight text-based alignment strategy which maps dense video captions to short summaries and enables task-related video-text embedding learning without visual supervision. Remarkably, without any fine-tuning beyond text, our method outperforms current methods, often by a substantial margin, achieving state-of-the-art results across common video retrieval benchmarks.
title VidVec: Unlocking Video MLLM Embeddings for Video-Text Retrieval
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2602.08099