Computer-aided Intelligent Diagnostics of Dynamic Perfusion CT Maps

This article presents novel automatic method for computer-aided diagnostics of dynamic perfusion CBF and CBV perfusion maps. The system enables the quantitative and quality analysis of visualized symptoms. The quantitative analysis consist detection, measure of potential lesions. The quality analysis is based on description and analysis of nature of invalid perfusion regions. The validation of the algorithm was performed on set of 75 perfusion images and compared to description done to each case by radiologist. Total error rate of the proposed solution was TER = 22.7% . Authors also present visualization module that is a perfect solution for the task of visualization input and output image data of the algorithm in the augmented reality technology.