Restoration of Scanned Document Image with Shape from Shading Method
This paper describes a method to restore document images from thick bound volumes. This restoration should remove two distortions: the photometric distortion and the geometric distortion, i.e., should remove the shade and the warp. During the process, the needed parameters are determined via observing the scanned document image from both the geometric and photometric views. When in the photometric view, the model and method are referred to the shape from shading algorithm. With the proposed method, the restored images coincides 99％ with the original document images and provides much better images for further dealings. It is useful in digital library and some related areas.
Yuefang Gao Fei Luo Yuge Xu Jianhui Yin
College of Automation South China Univ. of Tech. Guangzhou, 510640 Rockwell Automation Guangzhou, 510060
国际会议
南宁
英文
2007-07-20（万方平台首次上网日期，不代表论文的发表时间）