• Home
  • Search
  • Structured Multimodal Attentions for TextVQA.
  • Open Access IconOpen Access
  • Cite Icon58
  • https://doi.org/10.1109/tpami.2021.3132034Copy DOI Icon

Structured Multimodal Attentions for TextVQA.

  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Text based Visual Question Answering (TextVQA) is a recently raised challenge requiring models to read text in images and answer natural language questions by jointly reasoning over the question, textual information and visual content. Introduction of this new modality - Optical Character Recognition (OCR) tokens ushers in demanding reasoning requirements. Most of the state-of-the-art (SoTA) VQA methods fail when answer these questions because of three reasons: (1) poor text reading ability; (2) lack of textual-visual reasoning capacity; and (3) choosing discriminative answering mechanism over generative couterpart (although this has been further addressed by M4C). In this paper, we propose an end-to-end structured multimodal attention (SMA) neural network to mainly solve the first two issues above. SMA first uses a structural graph representation to encode the object-object, object-text and text-text relationships appearing in the image, and then designs a multimodal graph attention network to reason over it. Finally, the outputs from the above modules are processed by a global-local attentional answering module to produce an answer splicing together tokens from both OCR and general vocabulary iteratively by following M4C. Our proposed model outperforms the SoTA models on TextVQA dataset and two tasks of ST-VQA dataset among all models except pre-training based TAP. Demonstrating strong reasoning ability, it also won first place in TextVQA Challenge 2020. We extensively test different OCR methods on several reasoning models and investigate the impact of gradually increased OCR performance on TextVQA benchmark. With better OCR results, different models share dramatic improvement over the VQA accuracy, but our model benefits most blessed by strong textual-visual reasoning ability. To grant our method an upper bound and make a fair testing base available for further works, we also provide human-annotated ground-truth OCR annotations for the TextVQA dataset, which were not given in the original release. The code and ground-truth OCR annotations for the TextVQA dataset are available at https://github.com/ChenyuGAO-CS/SMA.

Similar Papers
  • Components

Supp1-3132034.pdf

  • Dec 17, 2021
  • Peng Wang
  • PDF
  • Research Article
  • Citations1

Development of text extraction technique using optical character recognition and morphological reconstruction to eliminate artifacts of image’s background

  • Feb 25, 2022
  • Eastern-European Journal of Enterprise Technologies
  • Wasan M Jwaid
  • Research Article
  • Citations29

Beyond OCR + VQA: Towards end-to-end reading and reasoning for robust and accurate textvqa

  • Jan 21, 2023
  • Pattern Recognition
  • Gangyan Zeng +7
  • Research Article
  • Citations2

Effects of Character Recognition with Shell Histogram Method Using Plate Characters

  • Dec 01, 2019
  • Politeknik Dergisi
  • Rukiye Uzun Arslan +2
  • PDF
  • Research Article
  • Citations73

Optical character recognition with neural networks and post-correction with finite state methods

  • Aug 20, 2020
  • International Journal on Document Analysis and Recognition (IJDAR)
  • Senka Drobac +1
  • PDF
  • Research Article
  • Citations22

Arabic Optical Character Recognition: A Review

  • Jan 01, 2023
  • Computer Modeling in Engineering & Sciences
  • Salah Alghyaline
  • Research Article
  • Citations300

Optical Character Recognition by Open source OCR Tool Tesseract: A Case Study

  • Oct 20, 2012
  • International Journal of Computer Applications
  • Chirag Patel +2
  • Conference Article

Application of Optical Character Recognizer Prototype and Convulsion Neural Network for Vehicle License Plate Detection

  • Apr 16, 2024
  • Engineering headway
  • Fuad Achmadi +3
  • Research Article
  • Citations3

User Collaboration in Mass Digitisation of Textual Materials

  • Jan 01, 2010
  • Asaf Tzadok +1
  • PDF
  • Research Article
  • Citations33

Artificially Intelligent Readers: An Adaptive Framework for Original Handwritten Numerical Digits Recognition with OCR Methods

  • May 26, 2023
  • Information
  • Parth Hasmukh Jain +5
  • Research Article
  • Citations6

HDPA: historical document processing and analysis framework

  • May 20, 2020
  • Evolving Systems
  • Ladislav Lenc +4
  • Conference Article
  • Citations5

Optimization of OCR in Detecting Research Proposal and Lecturer Community Service Documents using Thresholding Method

  • Sep 23, 2021
  • Aghus Sofwan +5
  • Conference Article

Warped document image correction method based on heterogeneous registration strategies

  • Mar 13, 2013
  • Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE
  • Lijing Tong +4
  • Book Chapter

Facing Uncertainty in Digitisation

  • Nov 22, 2011
  • Lukas Gander +5
  • Research Article
  • Citations2

An Accuracy Examination of OCR Tools

  • Oct 01, 2019
  • International Journal of Innovative Technology and Exploring Engineering
  • Jayesh Majumdar +15
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.