Improving the print quality of screenshots

This article considers a method of improving print quality for screenshots. The proposed method is based on detecting and vectorizing text areas on raster images. The main study is dedicated to smooth screen-text segmentation, determining background and text color, improving resolution, and recovering the contour of symbols and approximating them with Bezier curves. The proposed method is resistant to different colors, text sizes, and languages and makes it possible to obtain a sharp and correct text display for printing.

[1]  I. V. Safonov,et al.  Bottom-up document segmentation method based on textural features , 2011, Pattern Recognition and Image Analysis.

[2]  Andrew Y. Ng,et al.  Text Detection and Character Recognition in Scene Images with Unsupervised Feature Learning , 2011, 2011 International Conference on Document Analysis and Recognition.

[3]  Xiaoyi Jiang,et al.  Segmentation of Very Low Resolution Screen-Rendered Text , 2007, Ninth International Conference on Document Analysis and Recognition (ICDAR 2007).

[4]  Eli Packer,et al.  Detection and Segmentation of Antialiased Text in Screen Images , 2011, 2011 International Conference on Document Analysis and Recognition.

[5]  Farshideh Einsele-Aazami,et al.  Recognition of ultra low resolution, anti-aliased text with small font sizes , 2008 .

[6]  Sang Ho Kim,et al.  Generation of PDF with vector symbols from scanned document , 2013, Electronic Imaging.

[7]  Trevor Hastie,et al.  Multi-class AdaBoost ∗ , 2009 .

[8]  Alexander Vezhnevets,et al.  ‘ Modest AdaBoost ’ – Teaching AdaBoost to Generalize Better , 2005 .

[9]  Yoav Freund,et al.  Experiments with a New Boosting Algorithm , 1996, ICML.

[10]  Xiaoyi Jiang,et al.  Segmentation of Very Low Resolution Screen-Rendered Text , 2007 .