Andréa Weber and Jacques Facon
Abstract
From colored images of slides stained with the anti-PCNA-Cyclinemonoclonal antibody, we aim at obtaining the cellular proliferate rate as the ratio between the amount of proliferate cells and the total amount of cells in case of malignant non-Hodgkin lymphomas. Next step is to obtain a binary image from a previous process of color segmentation based on the HSI model. A filtering process for eliminating holes inside the cells, as well as for softening edges and eliminating noise is also necessary. Afterwards, we do the pattern recognition by using binary morphological tools. Grouped cells from different colonies make the segmentation process through colors difficult. Making use of other processes such as ultimate erosion markers and skeleton by influence zone (SKIZ) seems to be effective ways for obtaining the segmentation cells.