Reconstruction of Brain MRI Images and Detection of Tumour

Abstract

The field of medical visualization of organs are needed for accurate diagnosis and treatment of any disease. Brain tumour diagnosis and surgery also requires accurate3D visualization of the brain. Detection and 3D visualization of the brain and possibly tumours from MRI area computationally time consuming and error-prone task. The proposed system presents a 3D reconstruction model of the brain which greatly helps the radiologist to effectively diagnose and analyze brain. If the subject is in motion state or there occurs a movement while taking the scan, there might be distortions in the output scan image. In order to avoid such circumstances, it is better to reconstruct the 2D image into a 3D space as it is more effective. Thus, the quality of the scan image is much better. From such reconstructed images, the diseases associated with the foetus can be identified. By the help of more features the proposed method can be used for the diagnosis of diseases in the organs. By training the proposed system with more organs and features it can be used for the detection of various diseases from the reconstructed images. More than 200 images were used for the training and around 150 images were used for testing. Here the 2D image slices are undergone image preprocessing, image registration, and then reconstruction. For the image registration, the method used is discrete wavelet transform which is more suitable for medical imaging. The proposed system is applicable for the clinical practices.