An Self-daptive Chinese Layout Scanning Image Tilt Correction Method
A self-adaptive Chinese layout tilt correction method was proposed to address the issue of tilt in images generated after scanning Chinese layouts.Firstly,an image classification algorithm based on layout features was designed to fully explore the inherent connections between text data and quickly distinguish different types of text.Then,different strategies are adopted to adaptively estimate the tilt angle based on the distinguished text types,ultimately achieving tilt correction of Chinese version scanned images.In the experimental stage,compared with algorithms such as k-nearest neighbor clus-tering,Fourier transform,Hough transform,and projection contour analysis,the proposed method had shown the best comprehensive performance indicators.
万方数据
维普
