Title :
A reconfigurable neural network for pattern identification in CCD-based detectors
Author :
Alderighi, M. ; D´Angelo, S. ; Gummati, E. ; Sechi, G.R.
Author_Institution :
Ist. di Fisica Cosmica e Tecnologie Relative, CNR, Rome, Italy
Abstract :
The paper presents the definition of a neural image processing system to be used with intensified CCD-based detectors for photon counting applications. It consists of 25 neurons linked to a reconfigurable network. The use of neural network was proved to be a suitable choice for analysing relations among CCD pixels in a quite natural way. The adoption of a reconfigurable interconnection network allows to define and run different neural network topologies on the same hardware platform. Detector concepts are given and system layout is described in the paper
Keywords :
CCD image sensors; image recognition; neural nets; photon counting; reconfigurable architectures; CCD pixels; intensified CCD-based detectors; neural image processing; neural network topologies; pattern identification; photon counting applications; reconfigurable network; reconfigurable neural network; Charge coupled devices; Detectors; Face detection; Hardware; Intelligent networks; Multiprocessor interconnection networks; Network topology; Neural networks; Neurons; Phosphors;
Conference_Titel :
Instrumentation and Measurement Technology Conference, 1997. IMTC/97. Proceedings. Sensing, Processing, Networking., IEEE
Conference_Location :
Ottawa, Ont.
Print_ISBN :
0-7803-3747-6
DOI :
10.1109/IMTC.1997.612395
