Using neural network for liver detection in abdominal MRI images

MRI imaging is the one of useful abdominal imaging that the image parts are being demonstrated in high quality and clearness. Abdominal MRI images have been widely studied in the recent years as they are becoming an invaluable mean for abdominal organ investigation. In the field of medical image processing, some of the current interests are the automatic diagnosis of liver pathologies .The first and fundamental step in all these studies is the automatic liver segmentation that is still an open problem. In this paper we have presented new automatic system for liver segmentation from abdominal MRI images. This system includes two successive steps, pre-processing and liver image extraction algorithm. The pre-processing is applied for image enhancement (Edge preserved noise reduction) by using the mathematical morphology. After pre-processing, the abdominal MRI images are partitioned to some regions by using watershed algorithm. The feed forward neural network is used to liver features extraction in training stage. These features are used in liver recognition. Results show that this system recognizes the ridges of liver as well as physician liver extraction.