A New Paradigm for Geometric AccuracyPrediction in Medical Image Segmentation

Safety and accuracy are two important keywords when deal-ing with life-critical systems. In particular, in medical image processingthese two aspects have to be taken into careful attention since it represents the first step in the process that starts with the image acquisition and proceeds to the diagnosis step and therapy definition. Therefore it is important to analyze the possible inaccuracy sources that can be found in this step, since they will affect the accuracy of the whole system. In literature there are several techniques for the safety analysis and accuracy evaluation of complex systems, however most of the proposed approaches in the field of medical image processing only face the problem of defining different metrics that can accurately assess the accuracy of imaging processing techniques from a purely geometrical and quantitative point of view.In this paper we introduce a different approach to the problem of segmentation accuracy assessment based on the analysis of the critical aspects in the segmentation workflow that can affect the accuracy of the overall system, according to the specific clinical problem under investigation.We present a proof of feasibility of our approach by combining the use of Petri Nets for the modeling of the workflow of segmentation procedures in two different clinical scenarios: accuracy evaluation of manual segmentations performed by non-experts (skin tumors) and of a semi-automatic system (liver lesions).Results show that it is feasible to correlate the qualitative analysis withthe quantitative measures: in this way it is possible to predict the inaccuracy of the segmentation results and to optimize the different steps of the system before even acquiring and processing the data.