Classification of Malignant Lymphomas by Classifier Ensemble with Multiple Texture Features

摘要

Lymphoma is a cancer affecting lymph nodes. A reliable and precise classification of malignant lymphoma is essential for successful treatment. Current methods for classifying the malignancies rely on a variety of morphological, clinical and molecular variables. In spite of recent progress, there are still uncertainties in diagnosis. Automatic classification of images taken from slides with hematoxylin and eosin stained biopsy samples can allow more consistent and less labor-consuming diagnosis of this disease. In this paper, three well-known texture feature extraction methods including local binary patterns (LBP), Gabor filtering and Gray Level Coocurrence Matrix (GLCM) have been applied to efficiently represent the three types of malignancies, namely, Chronic Lym-photic Leukemia(CLL), Follicular Lymphoma (FL) cells, and Mantle Cell Lymphoma (MCL). Three classifiers of ^-Nearest Neighbor, multiple-layer per-ceptron and Support Vector Machine have been experimented and the simple classifier ensemble scheme majority-voting demonstrated obvious improvement in the classification performance.