Hyperspectral texture analysis for colon tissue biopsy classification

Diagnosis and cure of colon cancer can be improved by performing automated histopathological analysis of colon biopsy samples. Due to significant observational variation between pathologists in several histological features, there is a need for the development of automated, quantitative analysis techniques. This paper presents a promising automative technique for the classification of hyperspectral colon tissue biopsy images. The application of hyperspectral imaging techniques in medical image analysis is a new domain for researchers. The main advantage of using hyperspectral imaging is the increased spectral resolution and detailed subpixel information. The proposed classification algorithm is based on the subspace projection techniques particularly Support Vector Machines (SVM) with Gaussian Kernel. Dimensionality reduction and tissue segmentation is achieved by Independent Component Analysis (ICA) and k-means clustering. Morphological features, which describe the shape, orientation and other geometrical attributes, are extracted in first set of experiments. Grey level cooccurrence matrices are computed for the second set of experiments. The SVM with appropriate choice of parameters for its Gaussian kernel efficiently exploits the non linear boundary between the benign and malignant classes of the colon tissue biopsies