The study presents an automatic recognition method for parotid gland tumor. We used a convolution neural network to conduct the segmentation of parotid gland tumor and classifications of tumor types. We also designed eight combinations of various MRI contrasts to compare the results of recognition for parotid gland tumor. We compared results obtained using various combinations of MR images as the input of the convolutional neural network and found that diffusion-related parameters and contrast-enhanced T1 images played the primary role of the prediction accuracy.
This abstract and the presentation materials are available to members only; a login is required.