G. Le Certen

Cet article se propose de decrire les premieres etapes de developpement d'un Systeme Intelligent de Monitoring ou les notions de surveillance continue, de detection et d'identification en temps reel d'anomalies physiologiques, de formulation de diagnostics et de prescription de therapies, sont majeures. Notre volonte est de concevoir et de developper un Systeme de Monitoring Intelligent medical generique, independamment a la fois du domaine d'application medicale et de l'architecture materielle utilisee. La methodologie de specification temps reel SA/RT est en ce sens appliquee. A titre preliminaire a ce travail long terme, le prototype de bas niveau SIMBAD a ete developpe et est presente ici. Ce prototype est charge d'acquerir et de traiter les donnees signal brutes afin de transmettre aux autres modules du systeme de nonitoring, intelligent des donnees caracteristiques, exploitables pour la delivrance de diagnostic. En guise d'application, nous proposon...

Increased interest is presently given to three-dimensional (3-D) medical applications due to simultaneous emergence of high-resolution imaging systems and computer graphics techniques. This paper intends to present a survey of this field together with discussion and prospective views. The general framework is composed of six parts: data acquisition and preprocessing, polygonal object representation, voxel description, system architecture, medical applications, and prospects. Data base acquisition includes characterization of imaging modalities and medical specificity. Preprocessing schemes are directed to improvements (filtering), windowing, and spatial anisotropy (linear or spline interpolation). The two following sections are devoted to descriptions of the main object representations. Particular emphasis is given to optimal contour approximation, surface triangulation, mathematical surfaces on one hand, and cuberille and voxel representations on the other. Display capabilities--hi...