Noise and skull removal of brain magnetic resonance image using curvelet transform and mathematical morphology

Abstract

Improvement in detection and evaluation of brain tumour is an important task in medical field. MRI is a technology which enables the detection, diagnosis and evaluation. An automatic detection requires pre-processed image. Preprocessing makes the image segmentation more accurate. In preprocessing the noise removal, enhancement of image, artifact removal and skull stripping are carried out. Noise can be introduced by transmission errors and compression of the images. So it is essential to apply an efficient denoising technique to compensate such data corruption. Noise removal of an image still remains a challenge because noise removal introduces artifacts and causes blurring of the images. To remove noise from the MR image there are several techniques existing. Initially the noise is removed from the MR image using curve let transform. After the noise removal the skull stripping is carried out. MR image consists of both skull and brain tissue region. Usually the tumour will be found in brain region. So, for better evaluation the skull from MR image can be removed in skull stripping. This paper aims at providing the brain MR image segmentation process which makes the diagnosis and analysis of brain tumour easier.