Abstract
Scanning pages from books or magazines usually result in a digital image with warped
regions, which impacts the text quality and decreases OCR precision. To address this problem,
this work proposes a method using machine learning to estimate an unwarping matrix that can
fix the warping regions of the scanned page. Furthermore, this method allows the user
interaction with the final document, making it possible to adjust the level of unwarping
adjustment.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
INC, HP, "UNWARPING PAGES USING SEMANTIC SEGMENTATION INFORMATION AND DENSE IMAGE WARPING MATRIX", Technical Disclosure Commons, (February 10, 2021)
https://www.tdcommons.org/dpubs_series/4068