L. Raffo et al., DESIGN OF AN ASIP ARCHITECTURE FOR LOW-LEVEL VISUAL ELABORATIONS, IEEE transactions on very large scale integration (VLSI) systems, 5(1), 1997, pp. 145-153
We consider the design process of VLSI systems dedicated to the real-t
ime implementation of cooperative algorithms whose functionalities can
be characterized by multilayer ensembles of simple elements which int
eract locally, These algorithms are related, even though not exclusive
ly, to the implementation of various tasks in low-level machine vision
, The starting point in the design process is the formulation of the s
equential algorithm that computes the behavior of the system, Algorith
mic transformations are performed to expose the parallelism originally
present in the task, Given the description in terms of parallel loops
, we partition the system and organize it as a set of processing units
, The architectural structure of these units takes properly into accou
nt the algorithmic constraints on precision both in data representatio
n and computation, The program flow implemented by our programmable ar
chitectural solution (ASIP) is an iterative sequence of multiply-and-a
ccumulate operations performed in parallel. The programmability concer
ns both the structure/coefficients of the algorithm-depending on the s
pecific application-and its computational parameters, The architecture
's main blocks are described in VHDL and synthesized as a semi-custom
chip, using standard tools. Following this procedure, we designed an A
SIP core for performing real-time texture-based image segregation.