Random Blur Data Augmentation for Scene Text Recognition

In this paper, we propose to apply data augmentation approaches that provide more diverse training images, thus helping train robust deep models for the Scene Text Recognition (STR) task. The methods are Random Blur Region (RBR) and Units (RBUs). Specifically, first introduce RBR designed STR training, randomly selects a region sets pixels in with an average value. However, when provides various samples STR, it may make ambiguous reduce recognition accuracy. To address above problem, also RBUs divides blur into several units. Note of one unit share same way, can additional readable help models. Extensive experiments on datasets show achieve highly competitive performance. Besides, complementary commonly used techniques.