Nava Eslami - Dept. of computer engineering, Islamic Azad University, North Tehran Branch, Tehran, Iran.
Mohammad Mansour Riahi Kashani - Dept. of computer engineering, Islamic Azad University, North Tehran Branch, Tehran, Iran.
Amir Reza Radmard - Dept. of Radiology, Shariati Hospital, Tehran University of Medical Sciences, Tehran, Iran.
Ali Yoonessi - Neuroscience Department, School of Advanced Technologies in Medicine and National Brain Mapping Centre, Tehran University of Medical Sciences, Tehran, Iran.

Precise spleen extraction in Magnetic Resonance Imaging and volume calculation of this tissue with minimum user interference is an important topic of interest for diagnosis of spleen area diseases. In this paper a combined method for diagnosis of spleen in MR images and volume calculation of this tissue is presented. The algorithm is composed of three steps: preprocessing, segmentation and volume calculation. The preprocessing eliminates noise and improves the main edges and boundaries. Segmentation based on a snake and ratio of mean and variance of directional gradients is used for spleen extraction. Ratio of mean and variance of directional gradients is applied as external force (energy) in the snake. Finally, volume is calculated corresponding to the gap level between the slices. The proposed algorithm is tested on several T1-weighted MR images that are provided from patients of a hospital. The result of this algorithm is compared with the result of manual boundary tracing method and the average relative error percent is less than %5.0.

MRI Segmentation, Spleen Region Extraction, Mathematical Morphology, Snake, Ratio of mean and variance of directional gradients