EKTVQA: Generalized use of External Knowledge to empower Scene Text in Text-VQA

TL;DR

This paper introduces EKTVQA, a framework that leverages external knowledge to improve zero-shot scene text understanding in Text-VQA, enhancing reasoning, reducing bias, and improving answer accuracy.

Contribution

It presents a novel method for integrating external knowledge into multimodal transformers for Text-VQA, addressing zero-shot challenges and improving interpretability.

Findings

Achieves state-of-the-art results on three datasets.

Enhances answer entity correctness and multiword entity detection.

Reduces training data bias in scene text reasoning.

Abstract

The open-ended question answering task of Text-VQA often requires reading and reasoning about rarely seen or completely unseen scene-text content of an image. We address this zero-shot nature of the problem by proposing the generalized use of external knowledge to augment our understanding of the scene text. We design a framework to extract, validate, and reason with knowledge using a standard multimodal transformer for vision language understanding tasks. Through empirical evidence and qualitative results, we demonstrate how external knowledge can highlight instance-only cues and thus help deal with training data bias, improve answer entity type correctness, and detect multiword named entities. We generate results comparable to the state-of-the-art on three publicly available datasets, under the constraints of similar upstream OCR systems and training data.