Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chowdhury, Md Intisar, Aukkapinyo, Kittinun, Fujimura, Hiroshi, Woo, Joo Ann, Wasusatein, Wasu, Ghourabi, Fadoua
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2505.24371
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915412542750720
author Chowdhury, Md Intisar
Aukkapinyo, Kittinun
Fujimura, Hiroshi
Woo, Joo Ann
Wasusatein, Wasu
Ghourabi, Fadoua
author_facet Chowdhury, Md Intisar
Aukkapinyo, Kittinun
Fujimura, Hiroshi
Woo, Joo Ann
Wasusatein, Wasu
Ghourabi, Fadoua
contents In this paper, we propose a Grid-based Local and Global Area Transcription (Grid-LoGAT) system for Video Question Answering (VideoQA). The system operates in two phases. First, extracting text transcripts from video frames using a Vision-Language Model (VLM). Next, processing questions using these transcripts to generate answers through a Large Language Model (LLM). This design ensures image privacy by deploying the VLM on edge devices and the LLM in the cloud. To improve transcript quality, we propose grid-based visual prompting, which extracts intricate local details from each grid cell and integrates them with global information. Evaluation results show that Grid-LoGAT, using the open-source VLM (LLaVA-1.6-7B) and LLM (Llama-3.1-8B), outperforms state-of-the-art methods with similar baseline models on NExT-QA and STAR-QA datasets with an accuracy of 65.9% and 50.11% respectively. Additionally, our method surpasses the non-grid version by 24 points on localization-based questions we created using NExT-QA. (This paper is accepted by IEEE ICIP 2025.)
format Preprint
id arxiv_https___arxiv_org_abs_2505_24371
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Grid-LOGAT: Grid Based Local and Global Area Transcription for Video Question Answering
Chowdhury, Md Intisar
Aukkapinyo, Kittinun
Fujimura, Hiroshi
Woo, Joo Ann
Wasusatein, Wasu
Ghourabi, Fadoua
Computer Vision and Pattern Recognition
Artificial Intelligence
In this paper, we propose a Grid-based Local and Global Area Transcription (Grid-LoGAT) system for Video Question Answering (VideoQA). The system operates in two phases. First, extracting text transcripts from video frames using a Vision-Language Model (VLM). Next, processing questions using these transcripts to generate answers through a Large Language Model (LLM). This design ensures image privacy by deploying the VLM on edge devices and the LLM in the cloud. To improve transcript quality, we propose grid-based visual prompting, which extracts intricate local details from each grid cell and integrates them with global information. Evaluation results show that Grid-LoGAT, using the open-source VLM (LLaVA-1.6-7B) and LLM (Llama-3.1-8B), outperforms state-of-the-art methods with similar baseline models on NExT-QA and STAR-QA datasets with an accuracy of 65.9% and 50.11% respectively. Additionally, our method surpasses the non-grid version by 24 points on localization-based questions we created using NExT-QA. (This paper is accepted by IEEE ICIP 2025.)
title Grid-LOGAT: Grid Based Local and Global Area Transcription for Video Question Answering
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2505.24371