Training Protocol Matters: Towards Accurate Scene Text Recognition via Training Protocol Searching

TL;DR

This paper introduces a training protocol search method that optimizes hyper-parameters to significantly improve scene text recognition accuracy and speed, outperforming existing models.

Contribution

We propose a novel training protocol search algorithm using evolutionary optimization and proxy tasks, enhancing STR model performance and efficiency.

Findings

Improved recognition accuracy by 2.7% to 3.9% on mainstream STR models.

Achieved 2.1% higher accuracy with TRBA-Net compared to state-of-the-art.

Faster inference speeds on CPU and GPU with the searched training protocol.

Abstract

The development of scene text recognition (STR) in the era of deep learning has been mainly focused on novel architectures of STR models. However, training protocol (i.e., settings of the hyper-parameters involved in the training of STR models), which plays an equally important role in successfully training a good STR model, is under-explored for scene text recognition. In this work, we attempt to improve the accuracy of existing STR models by searching for optimal training protocol. Specifically, we develop a training protocol search algorithm, based on a newly designed search space and an efficient search algorithm using evolutionary optimization and proxy tasks. Experimental results show that our searched training protocol can improve the recognition accuracy of mainstream STR models by 2.7%~3.9%. In particular, with the searched training protocol, TRBA-Net achieves 2.1% higher accuracy than the state-of-the-art STR model (i.e., EFIFSTR), while the inference speed is 2.3x and 3.7x faster on CPU and GPU respectively. Extensive experiments are conducted to demonstrate the effectiveness of the proposed method and the generalization ability of the training protocol found by our search method. Code is available at https://github.com/VDIGPKU/STR_TPSearch.