Bidirectional Scene Text Recognition with a Single Decoder

Abstract

Scene Text Recognition (STR) is the problem of recognizing the correct word or character sequence in a cropped word image. To obtain more robust output sequences, the notion of bidirectional STR has been introduced. So far, bidirectional STRs have been implemented by using two separate decoders; one for left-to-right decoding and one for right-to-left. Having two separate decoders for almost the same task with the same output space is undesirable from a computational and optimization point of view. We introduce the bidirectional Scene Text Transformer (Bi-STET), a novel bidirectional STR method with a single decoder for bidirectional text decoding. With its single decoder, Bi-STET outperforms methods that apply bidirectional decoding by using two separate decoders while also being more efficient than those methods, Furthermore, we achieve or beat state-of-the-art (SOTA) methods on all STR benchmarks with Bi-STET. Finally, we provide analyses and insights into the performance of Bi-STET.

8 pages. In 24th European Conference on Artificial Intelligence

References in corpus (13)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• ADADELTA: An Adaptive Learning Rate Method
• Focusing Attention: Towards Accurate Text Recognition in Natural Images
• One Model To Learn Them All
• COCO-Text: Dataset and Benchmark for Text Detection and Recognition in Natural Images
• Synthetic Data for Text Localisation in Natural Images
• Deep Structured Output Learning for Unconstrained Text Recognition
• What Is Wrong With Scene Text Recognition Model Comparisons? Dataset and Model Analysis
• Sharp Nearby, Fuzzy Far Away: How Neural Language Models Use Context
• Reading Scene Text with Attention Convolutional Sequence Modeling
• ESIR: End-to-end Scene Text Recognition via Iterative Image Rectification
• NRTR: A No-Recurrence Sequence-to-Sequence Model For Scene Text Recognition
• Symmetry-constrained Rectification Network for Scene Text Recognition

Cited by in corpus (4)

• TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models
• Rethinking Text Line Recognition Models
• RobustScanner: Dynamically Enhancing Positional Clues for Robust Text Recognition
• Utilizing Resource-Rich Language Datasets for End-to-End Scene Text Recognition in Resource-Poor Languages