VQA-Aid: Visual Question Answering for Post-Disaster Damage Assessment and Analysis

TL;DR

This paper introduces HurMic-VQA, a new dataset for visual question answering to aid post-disaster damage assessment using UAV imagery, demonstrating its potential to improve real-time disaster response.

Contribution

The paper presents a novel VQA dataset, HurMic-VQA, specifically collected during a hurricane, and evaluates baseline models for post-disaster damage assessment.

Findings

HurMic-VQA dataset enables better scene understanding for disaster analysis.

Baseline VQA models show promising results on post-disaster imagery.

VQA systems can accelerate damage assessment and recovery efforts.

Abstract

Visual Question Answering system integrated with Unmanned Aerial Vehicle (UAV) has a lot of potentials to advance the post-disaster damage assessment purpose. Providing assistance to affected areas is highly dependent on real-time data assessment and analysis. Scope of the Visual Question Answering is to understand the scene and provide query related answer which certainly faster the recovery process after any disaster. In this work, we address the importance of \textit{visual question answering (VQA)} task for post-disaster damage assessment by presenting our recently developed VQA dataset called \textit{HurMic-VQA} collected during hurricane Michael, and comparing the performances of baseline VQA models.