
Proceedings Paper
A comparison of basic deinterlacing approaches for a computer assisted diagnosis approach of videoscope imagesFormat | Member Price | Non-Member Price |
---|---|---|
$17.00 | $21.00 |
Paper Abstract
In the near future, Computer Assisted Diagnosis (CAD) which is well known in the area of mammography
might be used to support clinical experts in the diagnosis of images derived from imaging modalities
such as endoscopy. In the recent past, a few first approaches for computer assisted endoscopy have been
presented already. These systems use a video signal as an input that is provided by the endoscopes
video processor. Despite the advent of high-definition systems most standard endoscopy systems today
still provide only analog video signals. These signals consist of interlaced images that can not be used
in a CAD approach without deinterlacing. Of course, there are many different deinterlacing approaches
known today. But most of them are specializations of some basic approaches. In this paper we present
four basic deinterlacing approaches. We have used a database of non-interlaced images which have been
degraded by artificial interlacing and afterwards processed by these approaches. The database contains
regions of interest (ROI) of clinical relevance for the diagnosis of abnormalities in the esophagus. We
compared the classification rates on these ROIs on the original images and after the deinterlacing. The
results show that the deinterlacing has an impact on the classification rates. The Bobbing approach
and the Motion Compensation approach achieved the best classification results in most cases.
Paper Details
Date Published: 9 March 2010
PDF: 8 pages
Proc. SPIE 7624, Medical Imaging 2010: Computer-Aided Diagnosis, 76243K (9 March 2010); doi: 10.1117/12.844366
Published in SPIE Proceedings Vol. 7624:
Medical Imaging 2010: Computer-Aided Diagnosis
Nico Karssemeijer; Ronald M. Summers, Editor(s)
PDF: 8 pages
Proc. SPIE 7624, Medical Imaging 2010: Computer-Aided Diagnosis, 76243K (9 March 2010); doi: 10.1117/12.844366
Show Author Affiliations
Andreas Kage, Fraunhofer-Institut für Integrierte Schaltungen (Germany)
Marcia Canto, The Johns Hopkins Univ. (United States)
Emmanuel Gorospe, The Johns Hopkins Univ. (United States)
Marcia Canto, The Johns Hopkins Univ. (United States)
Emmanuel Gorospe, The Johns Hopkins Univ. (United States)
Antonio Almario, The Johns Hopkins Univ. (United States)
Christian Münzenmayer, Fraunhofer-Institut für Integrierte Schaltungen (Germany)
Christian Münzenmayer, Fraunhofer-Institut für Integrierte Schaltungen (Germany)
Published in SPIE Proceedings Vol. 7624:
Medical Imaging 2010: Computer-Aided Diagnosis
Nico Karssemeijer; Ronald M. Summers, Editor(s)
© SPIE. Terms of Use
