Research Paper Explores Deep Learning and Brain Tumor Segmentation

Deep learning for brain tumor segmentation

A research paper by ADAPT researcher, Dr Malika Bendechache, highlights a new deep learning approach to identifying brain tumors that performs similar to human observers.  The paper titled “Brain tumor segmentation based on deep learning and an attention mechanism using MRI multi-modalities brain images”, was recently published in the Scientific Reports Journal.

In the paper, the researchers have developed a new brain tumor segmentation architecture that benefits from the characterization of the four MRI modalities. It means that each modality has unique characteristics to help the network efficiently distinguish between classes. By working only on a part of the brain image near the tumor tissue the authors showed that a CNN model, that is the most popular deep learning architecture, can reach performance close to human observers. 

Speaking about the research Dr Bendechache said: “In our method, after extracting the tumor’s expected area using a powerful preprocessing approach, those patches are selected to feed the network that their center is located inside this area. This leads to reducing the computational time and capability to make predictions fast for classifying the clinical image as it removes a large number of insignificant pixels off the image in the preprocessing step. Comprehensive experiments have indicated the effectiveness of the Distance-Wise Attention mechanism in our algorithm as well as the remarkable capacity of our entire model when compared with the state-of-the-art approaches.”