Abstract
The application of pattern recognition techniques to scintigrams is investigated. A preprocessing method which produces the 'curvature scintigram', and a boundary detection algorithm for feature extraction purposes are described. Clinically relevant parameters are then extracted from the bounded count rate scintigram and the curvature scintigram. For classification, the factor analysis of correspondence and the discriminant analysis are used. These procedures are applied to liver scintigrams of 47 patients who have been categorized by biopsy. As the results show, the curvature scintigram and the parameter extraction from the bounded count rate scintigrams improve the quality of the diagnosis and support the crucial distinction between diffuse 'patchy' structures and the presence of a few distinct lesions in scintigrams.
Vaknine, R;
Ammann, W W;
Lorenz, W J
[1]
- Deutsches Krebsforschungszentrum, Heidelberg (Germany, F.R.). Inst. fuer Nuklearmedizin
Citation Formats
Vaknine, R, Ammann, W W, and Lorenz, W J.
Investigation and application of pattern recognition techniques to medical picture data (scintigrams).
Germany: N. p.,
1977.
Web.
Vaknine, R, Ammann, W W, & Lorenz, W J.
Investigation and application of pattern recognition techniques to medical picture data (scintigrams).
Germany.
Vaknine, R, Ammann, W W, and Lorenz, W J.
1977.
"Investigation and application of pattern recognition techniques to medical picture data (scintigrams)."
Germany.
@misc{etde_7026852,
title = {Investigation and application of pattern recognition techniques to medical picture data (scintigrams)}
author = {Vaknine, R, Ammann, W W, and Lorenz, W J}
abstractNote = {The application of pattern recognition techniques to scintigrams is investigated. A preprocessing method which produces the 'curvature scintigram', and a boundary detection algorithm for feature extraction purposes are described. Clinically relevant parameters are then extracted from the bounded count rate scintigram and the curvature scintigram. For classification, the factor analysis of correspondence and the discriminant analysis are used. These procedures are applied to liver scintigrams of 47 patients who have been categorized by biopsy. As the results show, the curvature scintigram and the parameter extraction from the bounded count rate scintigrams improve the quality of the diagnosis and support the crucial distinction between diffuse 'patchy' structures and the presence of a few distinct lesions in scintigrams.}
journal = []
volume = {19:12}
journal type = {AC}
place = {Germany}
year = {1977}
month = {Dec}
}
title = {Investigation and application of pattern recognition techniques to medical picture data (scintigrams)}
author = {Vaknine, R, Ammann, W W, and Lorenz, W J}
abstractNote = {The application of pattern recognition techniques to scintigrams is investigated. A preprocessing method which produces the 'curvature scintigram', and a boundary detection algorithm for feature extraction purposes are described. Clinically relevant parameters are then extracted from the bounded count rate scintigram and the curvature scintigram. For classification, the factor analysis of correspondence and the discriminant analysis are used. These procedures are applied to liver scintigrams of 47 patients who have been categorized by biopsy. As the results show, the curvature scintigram and the parameter extraction from the bounded count rate scintigrams improve the quality of the diagnosis and support the crucial distinction between diffuse 'patchy' structures and the presence of a few distinct lesions in scintigrams.}
journal = []
volume = {19:12}
journal type = {AC}
place = {Germany}
year = {1977}
month = {Dec}
}