论文信息 - Compute-aided differentiation of focal liver lisease in MR imaging

Compute-aided differentiation of focal liver lisease in MR imaging

To aid radiologists for their decision of liver diseases in MR imaging, a method based on artificial neural network (ANN) is proposed. The input units correspond to the homogeneity and the intensities within region of interests selected from four MR images. In order to quantify the degree of liver cirrhosis, liver region is extracted from MR or CT images and divided into left and right parts. The ratio of volume in each part may differentiate a normal liver from cirrhosis. The result shows that our ANN based program may provide radiologists with educational use or referential opinion during their practical diagnostic procedure.

[1] H Fujita,et al. Application of artificial neural network to computer-aided diagnosis of coronary artery disease in myocardial SPECT bull's-eye images. , 1992, Journal of nuclear medicine : official publication, Society of Nuclear Medicine.

[2] Hiroshi Fujita,et al. Computerized classification of liver disease in MRI using an artificial neural network , 2001, SPIE Medical Imaging.

[3] K Arai,et al. Adenomatous hyperplastic nodules in the cirrhotic liver: differentiation from hepatocellular carcinoma with MR imaging. , 1989, Radiology.