Design And Application Of Ct Cerebral Tumor Perfusion

CT perfusion image which is one of the most hottest image processing technology in the field of image processing is taken to reveal tumors of physiological and pathological changes by the cell level and the level of microcirculation. There has not been a systematic theoretical research and its algorithm in domestic. Under the support of "National Digital Medical Image Equipment & Engineer Technology Research Center".We have used de-convolution algorithm in perfusion imaging in this paper and lied the theoretical foundation for cerebral tumor perfusion visualization software.Without any assumptions constraints for dual-chamber model which is selected to fit the tumor tissue. A digital de-convolution algorithm is introduced and realized to compute the impulse residue function R(t), many correlated mathematical theories such as Regularization matrix are used to keep the plate zone and attenuation zone of R(t).the SVD(Singular Value Decomposition) is used to reduce the high frequency and constrain the up of the tail of R(t), and the L-curve are used to decide the y coefficient and so on. At last, we got the curve of impulse residue function R (t) and hemodynamic parameters such as cerebral blood flow(CBF),cerebral blood volume(CBV),mean transit time (MTT) and the permeability coefficient (PS). The algorithm is tested by real CT cerebral tumors and medical image data, there is just a error of 4.6% between the results of the permeability coefficient with the real PS, and In different noise levels, the stability of the algorithm was verified.