Binarization of Color Document Image Based on Adversarial Generative Network and Discrete Wavelet Transform

Document binarization is an important task to separate the foreground text information in the document image from the background, which is generally applied to the digital archive of historical documents. This paper proposes to use the generative adversarial network for training with a small amount of data. In the first stage, discrete wavelets are used for image enhancement of four-channel images, and local binarization and global binarization are trained separately to obtain the final result in the second stage. The experimental results show that our proposed method has better performance than the classical algorithm on the DIBCO dataset.