In this paper an easy to use and still efficient parallel image processing framework is presented. The aim of this research is to provide a simply, portable and efficient programming interface for these algorithms. Performances is gained using a message passing distributed model on the JPVM platform. Examples and experimental results where shown. ). © : int + sliceImage § rows columns : int¨: boolean + getImageSlices § © : PFIPImageCollection + resetParameters § © + getParameters § © :

more »

... eterCollection + addParameter § p : PFIPParameter+ readUsrParameters § © + readImage § fileName : String¨: boolean + spawnWorkers § © + sendDataToWorkers § © + receiveResult § © + processResult § © + cleanup § © Fig. 2. PFIPMaster class (UML like representation) PFIPParameterCollection -Manage a collection of PFIP parameters, usually gather all parameters necessary in algorithm implemented by a PFIP worker The main purpose of them is to complete the functionality of PFIP framework.