Karyometric data by image analysis and their use in pathology
Klaus Voss,
Institute of Pathology, Humboldt University DDR - 1040 Berlin, Schumannstr. 20-21, GDR
Peter Hufnagl,
Institute of Pathology, Humboldt University DDR - 1040 Berlin, Schumannstr. 20-21, GDR
Hubert Martin,
Institute of Pathology, Humboldt University DDR - 1040 Berlin, Schumannstr. 20-21, GDR
Karl Roth,
Institute of Pathology, Humboldt University DDR - 1040 Berlin, Schumannstr. 20-21, GDR
Abstract
An overview is given on image analysis applications in histopathology. In comparison to automated cytology, there are only few practical efforts in this field. But they have shown that karyometric data yielded by image analysis give valuable information for description of histological specimens. Further in the paper, general principles of object segmentation, feature generation, object classification, and characterization of histological specimens by karyometric data are shortly discussed. These principles are realized in the author's image analysis system. Finally, a practical example is presented (brain tumours). The mean effective time for evaluating one specimen (out of 346 cases) is 5-10 minutes (300-1000 investigated nuclei).
Om dit artikel te citeren:
Klaus Voss, Peter Hufnagl, Hubert Martin & Karl Roth, «Karyometric data by image analysis and their use in pathology», Acta Stereologica [En ligne], Volume 2 (1983), Number 2 - Proceedings of the second symposium on morphometry in morphological diagnoses, september 7-9, 1983, Kuopio, Finland - Dec. 1983, 312-318 URL : https://popups.uliege.be/0351-580x/index.php?id=4204.