Signaltheoretische Analyse histologischer Daten im Ortsfrequenzraum

Abstract

In the present study, a semi-automatic segmentation and classification algorithm is proposed for the analysis of histological and cytological images. In view of the fact that histological and cytological images usually exhibit poor contrast and blurred outlines, classical segmentation algorithms often fail to detect relevant structures. A new algorithm for texture segmentation based on signal processing methods in combination with machine learning techniques was therefore developed.