Experimenting Grid Protocols to improve privacy preservation in efficient distributed image processing

In order to share interesting data located in various institutions, it may be necessary to ensure secure accesses and privacy preservation. For example, it is the case of biomedical images, since they are related to confidential information corresponding to real patients. This paper describes an approach to enable the use of PIMA(GE)(2) Lib, Parallel IMAGE processing GEnoa Library, in the the analysis of distributed medical images complying with privacy preservation policy imposed by the owner institutions. To achieve this goal, a set of specialized I/O functions was developed to enable secure remote data access and elaboration. exploiting the Grid infrastructure and without copy data from their original position. The work considers the use of the EGEE infrastructure and the GFAL API. The approach was experimented in the analysis of images obtained through the Tissue MicroArray technique; the tests gave positive results also in the execution time.