Your Search Results

Use this resource - and many more! - in your textbook!

AcademicPub holds over eight million pieces of educational content for you to mix-and-match your way.

Experience the freedom of customizing your course pack with AcademicPub!
Not an educator but still interested in using this content? No problem! Visit our provider's page to contact the publisher and get permission directly.

Inclusion of a priori information in segmentation of colon lumen for 3D virtual colonoscopy

By: Yang, F.; Liang, Z.; Viswambharan, A.; Li, H.; Li, J.; Hong, L.; Kaufman, A.; You, J.; Wax, M.;

1997 / IEEE / 0-7803-4258-5

Description

This item was taken from the IEEE Conference ' Inclusion of a priori information in segmentation of colon lumen for 3D virtual colonoscopy ' Segmentation of colon lumen from computed tomography (CT) is necessary for 3D virtual colonoscopy. Identifying the residual fluid retained inside the colon after colon-cleansing procedure is one of the major tasks for the segmentation. This work developed a computer algorithm to segment out the residual fluid after the image intensity of the fluid has been enhanced by ingesting an adequate amount of contrasting solution of Diatrizoate Meglumine and Diatrizoate Sodium. The algorithm models the image-intensity statistics across the field-of-view (FOV) as a mixture of Gaussian functionals and assumes a Markov random field (MRF) for the labels of the underling tissue distribution. The model parameters of the mixture are fitted to the image data by the maximum-likelihood estimator. In the fitting, the a priori known attenuation coefficients of air (inside the colon), soft tissue (fat) and muscle are included as the initial estimate. As iteration progresses, the initial estimate is tuned to fit into the image data. The optimal number of tissue types is determined by an information criterion. With the determined number of tissue types and the fitted model parameters of the mixture, each image pixel is classified obeying the assumption of MRF across the FOV. The algorithm was tested by acquired CT abdomen images. Its performance was very encouraging. The computing efficiency was significantly improved when the histogram of the image data was used.