Small Tumor-Aware Network for Brain MRI Image Segmentation

Brain tumor segmentation is a critical task in computer-aided diagnosis (CAD) systems for brain cancer detection because accurate tumor size, shape, and location are important for further tumor quantification and classification. However, segmenting small tumors in MRI images is challenging due to the speckle noise, varying tumor shapes and sizes among patients, and the existence of tumor-like image regions. Recently, deep learning-based approaches have achieved great success in biomedical image analysis, but current state-of-the-art approaches achieve poor performance for segmenting small brain tumors. In this paper, we propose a UNet++ model - Small Tumor-Aware Network, to accurately and robustly segment brain tumors. The Unet model with special fine tuning like Learning Rate, Loss Function and Hyperparameter Tuning. We compare proposed model with Base UNet against 3D Slicer segmented tumor. The results demonstrate that the proposed approach achieves the best overall performance and is on par with 3D Slicer manually segmented small tumor which is very tedious and tiresome.