# Structured Multimodal Attentions for TextVQA

arXiv: https://arxiv.org/abs/2006.00753 · HTML: https://mlsanity.com/paper/2006.00753/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 1 Jun 2020.

## Built on by (most cited first, 23 of 23)

- [GIT: A Generative Image-to-text Transformer for Vision and Language](https://mlsanity.com/paper/2205.14100/) (2022)
- [TextOCR: Towards large-scale end-to-end reasoning for arbitrary-shaped scene text](https://mlsanity.com/paper/2105.05486/) (2021)
- [LaTr: Layout-Aware Transformer for Scene-Text VQA](https://mlsanity.com/paper/2112.12494/) (2021)
- [Toward 3D Spatial Reasoning for Human-like Text-based Visual Question Answering](https://mlsanity.com/paper/2209.10326/) (2022)
- [DocTr: Document Transformer for Structured Information Extraction in Documents](https://mlsanity.com/paper/2307.07929/) (2023)
- [Towards Models that Can See and Read](https://mlsanity.com/paper/2301.07389/) (2023)
- [A First Look: Towards Explainable TextVQA Models via Visual and Textual Explanations](https://mlsanity.com/paper/2105.02626/) (2021)
- [Localize, Group, and Select: Boosting Text-VQA by Scene Text Modeling](https://mlsanity.com/paper/2108.08965/) (2021)
- [Universal Multimodal Representation for Language Understanding](https://mlsanity.com/paper/2301.03344/) (2023)
- [TAG: Boosting Text-VQA via Text-aware Visual Question-answer Generation](https://mlsanity.com/paper/2208.01813/) (2022)
- [Separate and Locate: Rethink the Text in Text-based Visual Question Answering](https://mlsanity.com/paper/2308.16383/) (2023)
- [Exploring Sparse Spatial Relation in Graph Inference for Text-Based VQA](https://mlsanity.com/paper/2310.09147/) (2023)
- [Locate Then Generate: Bridging Vision and Language with Bounding Box for Scene-Text VQA](https://mlsanity.com/paper/2304.01603/) (2023)
- [A Survey on Image-text Multimodal Models](https://mlsanity.com/paper/2309.15857/) (2023)
- [Multiple-Question Multiple-Answer Text-VQA](https://mlsanity.com/paper/2311.08622/) (2023)
- [Towards Escaping from Language Bias and OCR Error: Semantics-Centered Text Visual Question Answering](https://mlsanity.com/paper/2203.12929/) (2022)
- [SceneGATE: Scene-Graph based co-Attention networks for TExt visual question answering](https://mlsanity.com/paper/2212.08283/) (2022)
- [GRAFT: GRaPH and Table Reasoning for Textual Alignment -- A Benchmark for Structured Instruction Following and Visual Reasoning](https://mlsanity.com/paper/2508.15690/) (2025)
- [Making the V in Text-VQA Matter](https://mlsanity.com/paper/2308.00295/) (2023)
- [EKTVQA: Generalized use of External Knowledge to empower Scene Text in Text-VQA](https://mlsanity.com/paper/2108.09717/) (2021)
- [Graph Relation Transformer: Incorporating pairwise object features into the Transformer architecture](https://mlsanity.com/paper/2111.06075/) (2021)
- [Towards Multimodal Vision-Language Models Generating Non-Generic Text](https://mlsanity.com/paper/2207.04174/) (2022)
- [ViConsFormer: Constituting Meaningful Phrases of Scene Texts using Transformer-based Method in Vietnamese Text-based Visual Question Answering](https://mlsanity.com/paper/2410.14132/) (2024)

## Abstract (arXiv)

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