R
Renzo Perfetti
Researcher at University of Perugia
Publications - 56
Citations - 1436
Renzo Perfetti is an academic researcher from University of Perugia. The author has contributed to research in topics: Cellular neural network & Artificial neural network. The author has an hindex of 13, co-authored 56 publications receiving 1304 citations.
Papers
More filters
Journal ArticleDOI
Retinal Blood Vessel Segmentation Using Line Operators and Support Vector Classification
Elisa Ricci,Renzo Perfetti +1 more
TL;DR: In the framework of computer-aided diagnosis of eye diseases, retinal vessel segmentation based on line operators is proposed and two segmentation methods are considered.
Journal ArticleDOI
Cellular Neural Networks With Virtual Template Expansion for Retinal Vessel Segmentation
TL;DR: A retinal vessel segmentation method based on cellular neural networks (CNNs) is proposed, which is based on linear space-invariant 3times3 templates and can be realized using existing chip prototypes like the ACE16K.
Journal ArticleDOI
Rigorous Network and Full-Wave Electromagnetic Modeling of Wireless Power Transfer Links
TL;DR: In this paper, the authors make use of full-wave electromagnetic modeling of wireless power transfer links in order to derive the network characterization, eg, in terms of scattering or impedance matrix Once the latter is obtained, they show that network theory provides the appropriate matching impedances for achieving either maximum efficiency, maximum power on the load, or conjugate matching.
Journal ArticleDOI
Neural associative memory storing gray-coded gray-scale images
TL;DR: A neural associative memory storing gray-scale images stored in brain-state-in-a-box-type binary neural networks that guarantees asymptotic stability of the stored patterns, low computational cost, and control of the weights precision is presented.
Journal ArticleDOI
Analog neural network for support vector machine learning
Renzo Perfetti,Elisa Ricci +1 more
TL;DR: An analog neural network for support vector machine learning is proposed, based on a partially dual formulation of the quadratic programming problem, which results in a simpler circuit implementation with respect to existing neural solutions for the same application.