Separating Chinese Character from Noisy Background Using GAN

Abstract

Separating printed or handwritten characters from a noisy background is valuable for many applications including test paper autoscoring. The complex structure of Chinese characters makes it difficult to obtain the goal because of easy loss of fine details and overall structure in reconstructed characters. This paper proposes a method for separating Chinese characters based on generative adversarial network (GAN). We used ESRGAN as the basic network structure and applied dilated convolution and a novel loss function that improve the quality of reconstructed characters. Four popular Chinese fonts (Hei, Song, Kai, and Imitation Song) on real data collection were tested, and the proposed design was compared with other semantic segmentation approaches. The experimental results showed that the proposed method effectively separates Chinese characters from noisy background. In particular, our methods achieve better results in terms of Intersection over Union (IoU) and optical character recognition (OCR) accuracy.

Improved Ant Colony Algorithm Based on Binarization in Computer Text Recognition

“…The binarization algorithm divides the grayscale values of points on the entire image into two types at the threshold t .   0 0,1, 2, , Ct  represents the background area [23]. If the total pixels in the character area after grayscale processing are N , and each pixel has the highest grayscale level L , the grayscale size of the entire image is within   0, 1 L  .…”