Classifier Result Aggregation for Automatically Grading Histopathological Images
Cătălin Stoean, Daniel Lichtblau · 2017
The large amount of histopathological images that are produced in hospitals worldwide often request an overwhelming effort from the human pathology experts. In this respect, large efforts are made by scientists in various disciplines and especially in computer science to develop automatic procedures to distinguish between different grades of cancer based on the histopathological slides. The current work employs seven different machine learning approaches to classify the microscopical images from a real-world medical data set. Additionally, a combination of the results of the seven classifiers that gives a higher weight to a relatively new approach leads to classification results that are significantly better than those of the best performing algorithm of the considered collection.