abstract |
The task of segmenting cells in microscope images is a classical image analysis problem. The accurate segmentation may contribute to the development of successful feature extraction and classification. In this talk, we will look at two successive methods that have been used in our cell classification system. The first method applies a neural network trained with resilient backpropagation (RProp) to obtain a supervised classification of the cell nucleus. The second method uses this information to make a voronoi-based segmentation in the image.
|