Author(s)

Qifan Yang, Haomin Shao, Yi Li

Corresponding Author

Qifan Yang

ABSTRACT

In order to solve the problems of uneven illumination and low character quality in text image recognition, an image enhancement algorithm and a character recognition model based on convolution cyclic neural network are proposed in this paper. Among them, the image enhancement algorithm uses an improved tone mapping function considering local information to increase the visibility of text in dark areas. The method of background estimation and contrast compensation is used to solve the problem of uneven illumination of the image, and the connected domain method is used to locate the text in the image. The convolution and loop depth neural network model is built based on the text region, and the whole string in the image is taken as the recognition target. In this paper, 30 uneven illumination images are collected for experimental verification, and the experimental results show that the text recognition accuracy of the model in this scene is 98.29%.

KEYWORDS

Uneven illumination; local adaptive nonlinear filter; depth learning; text image recognition