Structured Multimodal Attentions for TextVQA

Abstract

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

arXiv commentwinner of TextVQA Challenge 2020, Accepted by IEEE Transactions on Pattern Analysis and Machine Intelligence

References in corpus (8)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• VQA: Visual Question Answering
• Rosetta: Large scale system for text detection and recognition in images
• COCO-Text: Dataset and Benchmark for Text Detection and Recognition in Natural Images
• FigureQA: An Annotated Figure Dataset for Visual Reasoning
• ICDAR2019 Robust Reading Challenge on Arbitrary-Shaped Text (RRC-ArT)
• TAP: Text-Aware Pre-training for Text-VQA and Text-Caption
• A First Look: Towards Explainable TextVQA Models via Visual and Textual Explanations

Cited by in corpus (6)

• TAP: Text-Aware Pre-training for Text-VQA and Text-Caption
• Towards Escaping from Language Bias and OCR Error: Semantics-Centered Text Visual Question Answering
• Finding the Evidence: Localization-aware Answer Prediction for Text Visual Question Answering
• Localize, Group, and Select: Boosting Text-VQA by Scene Text Modeling
• A First Look: Towards Explainable TextVQA Models via Visual and Textual Explanations
• Graph Relation Transformer: Incorporating pairwise object features into the Transformer architecture