Overview / Usage

Brain tumor segmentation is an important task in medical image processing. Early diagnosis of brain tumors plays an important role in improving treatment possibilities and increases the survival rate of the patients. Manual segmentation of the brain tumors for cancer diagnosis, from large amount of MRI images generated in clinical routine, is a difficult and time consuming task. There is a need for automatic brain tumor image segmentation. Recently, automatic segmentation using deep learning methods proved popular since these methods achieve the state-of-the-art results and can address this problem better than other methods. Deep learning methods with Conditional Random Fields can enable efficient processing and objective evaluation of the large amounts of MRI-based image data.

Methodology / Approach

Initially I was working on a CRF-based segmentation technique based on adaptive segmentation algorithms. The approach used CRF as a convenient means for introducing context or dependence among neighboring voxels. Currently, I am working on other graphical model approaches for brain tumor segmentation, such as Deep Belief Networks (DBNs) for manifold learning and Cascaded Neural Networks for learning the dependencies with multi-path CNNs.