A New Object-oriented Segmentation Algorithm based on CNNs - Part I: Edge Extraction
Giuseppe Grassi, Eugenio Di Sciascio, A.L. Grieco, Pietro Vecchio · 2005
By using the CNN paradigm, this paper and the companion one (Grasi et al., 2005) present a new object-oriented segmentation algorithm, which takes into account the hardware characteristics imposed by the CNNUM. In particular, by exploiting a rigorous model of the image contours, this paper focuses on CNN algorithms for edge extraction. Simulation results show that the approach provides more accurate edge extractions than the ones obtained by other CNN-based techniques.