Automatic spleen segmentation in MRI images using a combined neural network and recursive watershed transform

Accurate spleen segmentation in abdominal MRI images is one of the most important steps for computer aided spleen pathology diagnosis. The first and essential step for the diagnosis is the automatic spleen segmentation that is still an open problem. In this paper, we have proposed a new automatic algorithm for spleen area extraction in abdominal MRI images. The algorithm is fully automatic and contains several stages. The preprocessing stage is applied for required image enhancement. Then the abdominal MRI images are partitioned to different regions using combined recursive watershed transform and neural network. The feed forward neural network is trained and used for spleen features extraction. The features extracted using neural networks are used to monitor the quality of the output of watershed transform and adjusting required parameter automatically. The process of adjusting parameters is performed sequentially in several iterations. Experimental results showed the promise of the proposed algorithm.