templeblock / isbi2012

U-Net for ISBI Challenge 2012

Final test rand score: 0.979

The architecture was inspired by U-Net: Convolutional Networks for Biomedical Image Segmentation and implemented with Keras functional API.

Overview

Data

ISBI 2012 dataset: http://brainiac2.mit.edu/isbi_challenge/
The dataset is in folder data.

Model

src/models/structures.py

Training

training hyper-parameters:

• checkpoint_name: save_dir of tensorboard, set experiment name
• batch_size: batch size, integer multiple of GPU numbers
• prop_trn: proportion of training data 20./30
• prop_val: proportion of validation data
• montage_trn_shape: width and height of training data combination (5, 4) 5*4=20
• montage_val_shape: width and height of validation data combination (5, 2) 5*2=10
• early_stop_patience: 30
• aug: whether to use deformation in data augmentation
• steps: steps of each epoch
• random_split: whether to random split the 30 images data

Loss functions src/utils/model.py:

• weighted cross entropy
• focal loss
• dice loss
• tversky loss
• combinations

---

Usage

train.ipynb

train on the isbi training data

predict.ipynb

predict on validation or test data

data-augmentation.ipynb

check the results after different data augmentation

---

Acknowlegement

alexklibisz work