A Unified Geometric Model for Virtual Slide Image Processing and Classification

In this paper, we use the framework of partial difference equations on weighted graphs as a methodology to address the problem of computer-aided cytology. First, introduced to perform image smoothing and filtering, this framework has recently been extended to address segmentation and semi-supervised clustering of any discrete domain that can be represented by a graph of arbitrary topology. In particular, this framework unifies methods from image processing and machine learning communities within the same formulation. We demonstrate that this method can also be used effectively to unify preprocessing, image segmentation, and data classification for both Feulgen- and Papanicolaou-stained virtual slide processing as well as 3-D confocal microscopy. For evaluation we compare this approach to state-of-the-art algorithms for segmentation and classification.