Authors: Ning Lu*, Wenwen Yu*, Xianbiao Qi*, Yihao Chen, Rong Xiao

Affiliation: Ping An Property & Casualty Insurance Co

Description: We propose a novel approach MASTER: Multi-Aspect Non-local Network for Scene Text Recognition, a self-attention based scene text recognizer. It consists of two key modules, a Multi-Aspect Global Context Attention (GCAttention) based encoder and a Transformer based decoder. The proposed MASTER owns three advantages: (1) The model can both learn input-output attention and self-attention which encodes feature-feature and target-target relationships inside the encoder and decoder. (2) Experiments demonstrate that the proposed method is more robust to spatial distortion. (3) The training process of the proposed method is highly parallel and efficient. Experiments on standard benchmarks demonstrate it can achieve the state-of-the-art performances regarding both efficiency and recognition accuracy.

Authors: Jeonghun Baek, Junyeop Lee, Sungrae Park, Moonbin Yim, Seonghyeon Kim, Hwalsuk Lee

Description: We used Thin-plate-spline (TPS) based Spatial transformer network (STN) which normalizes the input text images, ResNet based feature extractor, BiLSTM, and attention mechanism.
This model was developed based on the analysis of scene text recognition modules.
See our paper and source code.

Authors: Hui Li, Peng Wang, Chunhua Shen, Guyu Zhang

Description: We propose an easy-to-implement strong baseline for irregular scene text recognition, using off- the-shelf neural network components and only word-level annotations. It is composed of a 31-layer ResNet, an LSTM- based encoder-decoder framework and a 2-dimensional attention module. Despite its simplicity, the proposed method is robust. It achieves state-of-the-art performance on irregular text recognition benchmarks and comparable results on regular text datasets.