Paper Info
Title
Fast and Robust Perspective Rectification of Document Images on a Smartphone
Abstract
This paper presents a perspective rectification framework for mobile device that is fast and robust to recovering the fronto-parallel view of perspectively distorted document images. The conventional approaches is too heavy to be implemented on a mobile device. In addition, they fails to reject the false case of the perspective rectification. To ameliorate such problems, the proposed framework is designed to pursue a fast and robust algorithm to detect horizontal and vertical vanishing points efficiently and robustly. Then, perspective rectification is performed using both horizontal and vertical vanishing points. In addition, the proposed framework has an adaptive scheme to detect the false case of the perspective rectification and skip the procedure without using the vertical vanishing point. Note that, the proposed framework is designed for consumer application so that bad results are rejected before they are shown for users. We demonstrate the performance of the proposed framework on various challenging examples to confirm that the proposed system is fast and robust in rectifying the perspectively distorted images.
Year
DOI
Venue
2014
10.1109/CVPRW.2014.37
CVPR Workshops
Keywords
Field
DocType
mobile device,distortion,smartphone,document image perspective rectification,feature extraction,vertical vanishing points detection,smart phones,document image processing,horizontal vanishing points detection,perspectively distorted document images
Computer vision,Horizontal and vertical,Rectification,Computer science,Mobile device,Artificial intelligence,Vanishing point
Conference
ISSN
Citations 
PageRank 
2160-7508
1
0.34
References 
Authors
3
4
Name
Order
Citations
PageRank
Williem1145.10
Christian Simon292.89
Sungdae Cho317715.35
In Kyu Park431635.97