The kind of cells considered in this application are Hep-2 cells, which get used for the identification of antinuclear autoantibodies (ANA). Hep-2 cells allow recognition of over 30 different nuclear and cytoplasmic patterns, which are given by upwards of 100 different autoantibodies. The identification of the patterns is usually done manually by a human inspecting the slides with a microscope. In the paper, we present first results on image analysis, feature extraction and classification. Starting from a knowledge acquisition process with a human operator, we developed a image analysis and feature extraction algorithm. A data set containing 112 features for each entry was set up and given to machine learning techniques to find out the relevant features among this large feature set and to construct the structure of the classifier. The classifier was evaluated by a cross validation method. The results are good and show the feasibility of an automatic inspection system.
[1]
Casimir A. Kulikowski,et al.
Computer Systems That Learn: Classification and Prediction Methods from Statistics, Neural Nets, Machine Learning and Expert Systems
,
1990
.
[2]
Heinrich Niemann.
Pattern Analysis and Understanding
,
1990
.
[3]
N. Otsu.
A threshold selection method from gray level histograms
,
1979
.
[4]
Petra Perner,et al.
Knowledge Acquisition by Symbolic Decision Tree Induction for Interpretation of Digital Images in Radiology
,
1996,
SSPR.
[5]
Petra Perner,et al.
A knowledge-based image-inspection system for automatic defect recognition, classification, and process diagnosis
,
1994,
Machine Vision and Applications.