Full Processing on Luna16 Challange (All data resampled to 1mm x 0.7mm x 0.7mm)
- First all volume data resampled to 0.7mm x 0.7mm x 0.7mm in order to make training better
- Mini-Batch is used to overcome class inbalance (Keras (with tensorflow backend) is used)