Saved in:
Bibliographic Details
Main Authors: Shen, Zhixuan, Luo, Haonan, Li, Sijia, Li, Tianrui
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.09288
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910367181963264
author Shen, Zhixuan
Luo, Haonan
Li, Sijia
Li, Tianrui
author_facet Shen, Zhixuan
Luo, Haonan
Li, Sijia
Li, Tianrui
contents Scene-Text Visual Question Answering (ST-VQA) aims to understand scene text in images and answer questions related to the text content. Most existing methods heavily rely on the accuracy of Optical Character Recognition (OCR) systems, and aggressive fine-tuning based on limited spatial location information and erroneous OCR text information often leads to inevitable overfitting. In this paper, we propose a multimodal adversarial training architecture with spatial awareness capabilities. Specifically, we introduce an Adversarial OCR Enhancement (AOE) module, which leverages adversarial training in the embedding space of OCR modality to enhance fault-tolerant representation of OCR texts, thereby reducing noise caused by OCR errors. Simultaneously, We add a Spatial-Aware Self-Attention (SASA) mechanism to help the model better capture the spatial relationships among OCR tokens. Various experiments demonstrate that our method achieves significant performance improvements on both the ST-VQA and TextVQA datasets and provides a novel paradigm for multimodal adversarial training.
format Preprint
id arxiv_https___arxiv_org_abs_2403_09288
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Adversarial Training with OCR Modality Perturbation for Scene-Text Visual Question Answering
Shen, Zhixuan
Luo, Haonan
Li, Sijia
Li, Tianrui
Computer Vision and Pattern Recognition
Artificial Intelligence
Scene-Text Visual Question Answering (ST-VQA) aims to understand scene text in images and answer questions related to the text content. Most existing methods heavily rely on the accuracy of Optical Character Recognition (OCR) systems, and aggressive fine-tuning based on limited spatial location information and erroneous OCR text information often leads to inevitable overfitting. In this paper, we propose a multimodal adversarial training architecture with spatial awareness capabilities. Specifically, we introduce an Adversarial OCR Enhancement (AOE) module, which leverages adversarial training in the embedding space of OCR modality to enhance fault-tolerant representation of OCR texts, thereby reducing noise caused by OCR errors. Simultaneously, We add a Spatial-Aware Self-Attention (SASA) mechanism to help the model better capture the spatial relationships among OCR tokens. Various experiments demonstrate that our method achieves significant performance improvements on both the ST-VQA and TextVQA datasets and provides a novel paradigm for multimodal adversarial training.
title Adversarial Training with OCR Modality Perturbation for Scene-Text Visual Question Answering
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2403.09288