Abstract
Scene Text Recognition (STR) remains a challenging task due to complex visual appearances and limited semantic priors. We propose TEACH, a novel training paradigm that injects ground-truth text into the model as auxiliary input and progressively reduces its influence during training. By encoding target labels into the embedding space and applying loss-aware masking, TEACH simulates a curriculum learning process that guides the model from label-dependent learning to fully visual recognition. Unlike language model-based approaches, TEACH requires no external pretraining and introduces no inference overhead. It is model-agnostic and can be seamlessly integrated into existing encoder-decoder frameworks. Extensive experiments across multiple public benchmarks show that models trained with TEACH achieve consistently improved accuracy, especially under challenging conditions, validating its robustness and general applicability.
Abstract (translated)
场景文本识别（STR）由于复杂的视觉外观和有限的语义先验，仍是一项具有挑战性的任务。我们提出了一种新的训练范式TEACH，该方法在模型中注入真实的文本作为辅助输入，并在训练过程中逐步减少其影响。通过将目标标签编码到嵌入空间并通过损失感知掩码进行处理，TEACH模拟了渐进学习过程，引导模型从依赖于标签的学习转向完全基于视觉的识别。与基于语言模型的方法不同，TEACH不需要外部预训练并且不会增加推理开销。该方法是模型不可知的，并且可以无缝集成到现有的编码器-解码器框架中。 在多个公共基准测试中的广泛实验表明，使用TEACH进行训练的模型在各种条件下的一致性准确度得到了提高，尤其是在挑战性的环境中表现尤为出色，这验证了其鲁棒性和通用适用性。
URL
https://arxiv.org/abs/2508.01153
