DocumentCode
1474136
Title
Digital imaging colposcopy: corrected area measurements using shape-from-shading
Author
Craine, Brian L. ; Craine, Eric R. ; O´Toole, Cynthia J. ; Ji, Qiang
Author_Institution
Western Res. Co. Inc., Tucson, AZ, USA
Volume
17
Issue
6
fYear
1998
Firstpage
1003
Lastpage
1010
Abstract
The quantitative measurement of areas on the cervix is of interest to researchers studying the natural history of human papilloma viral lesions. Measurement of areas from images obtained through a colposcope are, however, inherently in error due to the image being a two-dimensional projection of a three-dimensional object. The ability to correct for these errors through use of digital imaging colposcopy and a practical application of a shape from shading algorithm was developed in this study. The shape from shading technique requires empirical measurement of the relationship between observed light intensity and the viewing angle (referred to as a reflection map). It was found that a population mean reflection map provided a correction that was about as accurate as using an individual´s own reflection map (making it unnecessary to measure a map for each exam). Digital red filtering of the images increased accuracy and precision of measurement.
Keywords
area measurement; gynaecology; measurement errors; medical image processing; optical images; cervix; corrected area measurements; digital imaging colposcopy; digital red filtering; human papilloma viral lesions; measurement precision; medical diagnostic imaging; quantitative measurement; reflection map; shape-from-shading; three-dimensional object; two-dimensional projection; Area measurement; Digital filters; Digital images; Error correction; Filtering; History; Humans; Lesions; Optical reflection; Shape measurement; Adolescent; Adult; Algorithms; Cervix Uteri; Colposcopes; Colposcopy; Diagnosis, Computer-Assisted; Female; Humans; Least-Squares Analysis; Microcomputers; Middle Aged; Papillomaviridae; Papillomavirus Infections; Reproducibility of Results; Tumor Virus Infections; Uterine Cervical Diseases; Video Recording;
fLanguage
English
Journal_Title
Medical Imaging, IEEE Transactions on
Publisher
ieee
ISSN
0278-0062
Type
jour
DOI
10.1109/42.746633
Filename
746633
Link To Document