DocumentCode :
1016941
Title :
Performance Evaluation and Benchmarking of Six-Page Segmentation Algorithms
Author :
Shafait, Faisal ; Keysers, Daniel ; Breuel, Thomas M.
Author_Institution :
Image Understanding & Pattern Recognition Res. Group, German Res. Center for Artificial Intell., Kaiserslautern
Volume :
30
Issue :
6
fYear :
2008
fDate :
6/1/2008 12:00:00 AM
Firstpage :
941
Lastpage :
954
Abstract :
Informative benchmarks are crucial for optimizing the page segmentation step of an OCR system, frequently the performance limiting step for overall OCR system performance. We show that current evaluation scores are insufficient for diagnosing specific errors in page segmentation and fail to identify some classes of serious segmentation errors altogether. This paper introduces a vectorial score that is sensitive to, and identifies, the most important classes of segmentation errors (over, under, and mis-segmentation) and what page components (lines, blocks, etc.) are affected. Unlike previous schemes, our evaluation method has a canonical representation of ground-truth data and guarantees pixel-accurate evaluation results for arbitrary region shapes. We present the results of evaluating widely used segmentation algorithms (x-y cut, smearing, whitespace analysis, constrained text-line finding, docstrum, and Voronoi) on the UW-III database and demonstrate that the new evaluation scheme permits the identification of several specific flaws in individual segmentation methods.
Keywords :
document image processing; image segmentation; optical character recognition; OCR system; UW-III database; Voronoi; arbitrary region shapes; benchmarking; constrained text-line finding; docstrum; ground-truth data canonical representation; informative benchmarks; performance evaluation; pixel-accurate evaluation; segmentation errors; six-page segmentation algorithms; smearing; whitespace analysis; x-y cut; Document analysis; Optical character recognition; Algorithms; Artificial Intelligence; Automatic Data Processing; Benchmarking; Computer Graphics; Documentation; Image Enhancement; Image Interpretation, Computer-Assisted; Information Storage and Retrieval; Models, Statistical; Numerical Analysis, Computer-Assisted; Pattern Recognition, Automated; Reproducibility of Results; Sensitivity and Specificity; Signal Processing, Computer-Assisted; Subtraction Technique; User-Computer Interface;
fLanguage :
English
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
Publisher :
ieee
ISSN :
0162-8828
Type :
jour
DOI :
10.1109/TPAMI.2007.70837
Filename :
4407728
Link To Document :
بازگشت