A New Heterogeneous and Reconfigurable Architecture for Image Analysis

In the paper a new architecture for image analysis: HERMIA (Heterogeneous and Reconfigurable Machine for Image Analysis) is presented. It has been developed at the University of Palermo, inside the Progetto Finalizzato of the Italian Council of Researches (CNR): Sistemi informatici e Calcolo Parallelo. The architecture of the HERMIA-machine is reconfigurable, moreover the integration of heterogeneous module, oriented to the solution of specific problems, allows to solve complex problems by search of optimal strategies. Signal processing units allows the user to handle and integrate multi-sensors signals (from video, scanner, music recorder). Here the general architecture, the hardware implementation, and the iconic environment, of the HERMIA-machine are described. The preliminary system evaluation and foreseen applications are discussed.

[1]  Gaetano Gerardi,et al.  Integration of a structural features-based preclassifier and a man-machine interactive classifier for a fast multi-stroke character recognition , 1992, Proceedings., 11th IAPR International Conference on Pattern Recognition. Vol. IV. Conference D: Architectures for Vision and Pattern Recognition,.

[2]  Franco Chiavetta,et al.  HERMIA: An Heterogeneous and Reconfigurable Machine for Image Analysis , 1990, MVA.

[3]  Vito Di Gesù.  A high level language for pyramidal architectures , 1986 .

[4]  Kenneth E. Batcher,et al.  Design of a Massively Parallel Processor , 1980, IEEE Transactions on Computers.

[5]  Gaetano Gerardi,et al.  The Programmable and Configurable Low Level Vision Unit of the HERMIA Machine , 1992, MVA.

[6]  Michael J. Quinn,et al.  Designing Efficient Algorithms for Parallel Computers , 1987 .

[7]  O. Catalano,et al.  The PAPIA Image Analysis System , 1986, Other Conferences.

[8]  Virginio Cantoni,et al.  The PAPIA system , 1991, J. VLSI Signal Process..

[9]  W. Daniel Hillis,et al.  The connection machine , 1985 .