Non-rigid Registration of Brain MR Images for Image Guided Neurosurgery Using Cloud Computing

Image guided Surgery (IGS) has become familiar on enhancing it with computerized technologies which makes processing much easier. These computerized systems are highly sophisticated and improvised which orients the surgeon to perform the surgery with three dimensional images of the patients MR images. This work presents the Non Rigid Registration of Brain MRI with the help of Cloud computing. There exist geometric ambiguities among individuals due to the varying shape of the cortex and this is a challenging task when it comes to registration process. Intensity based registration is not alone enough to tackle the issues hence anatomical knowledge is highly important. Segmentation and labelling of cortex sulci with non-parametric approach that enables capturing of its shape and topology. The intensity and feature points are matched using the registration energy. A linear combination of compact smooth linear field and branched basic function is formed. The addition of the sulci from the upper border till its bottom part makes the registration process even more efficient. Cloud computing has been used in two scenarios, Firstly, the complete workflow of registration is enabled at the cloud to handle hypothetical computations and obtain precise results. Secondly, storage and computational nodes are used to handle large amount number of images nearly 7TB. The Deformable registration is evaluated and its feasibility has been analysed using cloud resources which resulted in timely execution of complex components during registration. The result obtained has proved that Cloud Computing provides practical and cost effective support for telemedicine while performing Image guided Surgery. The accuracy level has been improved upto 62% while performing Non Rigid Registration.