Text Recognition in the Wild: A Survey

摘要: The history of text can be traced back over thousands years. Rich and precise semantic information carried by is important in a wide range vision-based application scenarios. Therefore, recognition natural scenes has been an active research field computer vision pattern recognition. In recent years, with the rise development deep learning, numerous methods have shown promising terms innovation, practicality, efficiency. This paper aims to (1) summarize fundamental problems state-of-the-art associated scene recognition; (2) introduce new insights ideas; (3) provide comprehensive review publicly available resources; (4) point out directions for future work. summary, this literature attempts present entire picture It provides reference people entering field, could helpful inspire research. Related resources are at our Github repository: https URL.