Survey of Text Recognition Processes in Natural Scenes
Download as PDF
DOI: 10.23977/cii2019.05
Author(s)
Minghao Diao, Yanjun Yin
Corresponding Author
Minghao Diao
ABSTRACT
With the rapid development of Internet and artificial intelligence technology and the maturity of OCR technology, more software applications with OCR technology have entered people's daily lives. With the continuous development of deep learning technology in the field of vision, OCR technology has also broken through the bottleneck of the traditional technology framework. In recent years, a variety of detection algorithms such as anchor-based, pixel-based and pixel-anchor have emerged, and the use of machine recognition instead of manual entry has become a research hotspot. In this paper, the general research process of OCR in natural scenes, namely image preprocessing, text line location detection and text recognition, is reviewed. The problems of OCR text recognition and future research directions are expounded.
KEYWORDS
OCR, detection algorithm, recognition technology