3D Brain Segmentation Using 2D Fully Convolutional Neural Networks

Objectives

The goal of our experiments is to perform the brain extraction, also known as skull-stripping, using a 2D U-Shape fully convolutional neural network (CNN), which we chose the U-NET architecture.

Workflow

Data

The dataset used is the CC-359 which consists of T1 volumes acquired in 359 subjects on scanners from three different vendors (GE, Philips, and Siemens) and at two magnetic field strengths (1.5 T and 3 T). The dataset has twelve volumes with manual segmentations that will be used to evaluate the proposed method.

Tools

In this project, we basically use python to code and ITK-SNAP to see the 3D volume.

Libraries

NumPy, Scikit-Learn, Matplotlib, Keras

References

[1] Kleesiek, J. et al. NeuroImage, 129:460–469, 2016. [2] Warfield, S. K. et al. IEEE Tran. on Med. Imaging, 23(7):903–921, 2004. [3] Ronneberger, O. et al, MICCAI, Springer, 9351: 234-241, 2015.


In [ ]: