Visual exploration of pathology images by a discrete wavelet transform preprocessed locally linear embedding

Abstract

The information content of large collections of histopathological images can be explored utilizing computer-based techniques that can help the user to explore the similarity between different brain tumor types. To visually inspect the degree of similarity between different tumors, we propose a combined approach based on the Discrete Wavelet Transform (DWT) and Locally Linear Embedding (LLE). The former is employed as a preprocessing utility, the latter achieves the dimensional reduction required for visualization.