Iterative Answer Prediction With Pointer-Augmented Multimodal Transformers for TextVQA

• Main Authors: Hu, Ronghang, Singh, Amanpreet, Darrell, Trevor, Rohrbach, Marcus
• Format: Conference Proceeding
• Language: English
• Subjects: Feature extraction; Iterative decoding; Optical character recognition software; Predictive models; Task analysis; Visualization; Vocabulary
• ISSN: 2575-7075
• DOI: 10.1109/CVPR42600.2020.01001

Many visual scenes contain text that carries crucial information, and it is thus essential to understand text in images for downstream reasoning tasks. For example, a deep water label on a warning sign warns people about the danger in the scene. Recent work has explored the TextVQA task that requires reading and understanding text in images to answer a question. However, existing approaches for TextVQA are mostly based on custom pairwise fusion mechanisms between a pair of two modalities and are restricted to a single prediction step by casting TextVQA as a classification task. In this work, we propose a novel model for the TextVQA task based on a multimodal transformer architecture accompanied by a rich representation for text in images. Our model naturally fuses different modalities homogeneously by embedding them into a common semantic space where self-attention is applied to model inter- and intra- modality context. Furthermore, it enables iterative answer decoding with a dynamic pointer network, allowing the model to form an answer through multi-step prediction instead of one-step classification. Our model outperforms existing approaches on three benchmark datasets for the TextVQA task by a large margin.