数字内容技术与服务研究中心知识库

Abstract—Text is an important clue of visual recognition, and character recognition in natural scenes is one of the main research directions of computer vision. However, due to the similarity of characters, various styles, blurred images, complex background, uneven illumination, etc., there are still great difficulties to text recognition. In recent years, with the vigorous development of deep learning, text recognition has achieved promising performance. This paper first summarizes the challenges in text recognition; secondly, it introduces the state-of-the-art text recognition methods based on CTC and attention mechanism; finally, this paper also presents insights into potential research directions of text recognition. In summary, this literature review attempts to present the recent picture in the field of scene text recognition. It provides a reference for people entering this field and may help inspire future research.