Knowledge-based femur detection in conventional radiographs of the pelvis

Authors:
Roland Pilgram;Claudia Walch;Michael Blauth;Werner Jaschke;Rainer Schubert;Volker Kuhn
Affiliations:
Institute for Applied Systems Research and Development, University for Health Sciences, Medical Informatics and Technology (UMIT), Austria and Institute of Biomedical Image Analysis, University fo ...;Department of Radiology, Medical University of Innsbruck, Austria;Department of Trauma Surgery and Sport Traumatology, Medical University of Innsbruck, Austria;Department of Radiology, Medical University of Innsbruck, Austria;Institute of Biomedical Image Analysis, University for Health Sciences, Medical Informatics and Technology (UMIT), Austria;Department of Trauma Surgery and Sport Traumatology, Medical University of Innsbruck, Austria
Venue:
Computers in Biology and Medicine
Year:
2008

Citing 3
Cited 0

A Computational Approach to Edge Detection

IEEE Transactions on Pattern Analysis and Machine Intelligence
Active shape models—their training and application

Computer Vision and Image Understanding
The Use of Active Shape Models for Locating Structures in Medical Images

IPMI '93 Proceedings of the 13th International Conference on Information Processing in Medical Imaging

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper we present a knowledge-based femur detection algorithm. The algorithm uses femur corpus constraints, Canny edge detection and Hough lines. For optimal femur template placement in the local area we use cross-correlation. The segmentation itself is done with an optimized active shape modeling technique. Using the knowledge-based technique we have located 95% of the femur shapes of N=117 X-rays. From those 83% of the target femur shapes have been segmented successfully (point-to-point error: ~14 pixels, point-to-boundary error = ~9 pixels).