Deep learning applied to medical images course

Code and exercises for the Deep learning applied to medical images course

• 00_intro_to_keras
• 01_net_from_scrach_blood
• 02_data_augmentation_blood
• 03_transfer_learning_cats_dogs
• 04_object_detection_blood

IMPORTANT

The class will work all on cloud. Exercises and environment instructions are provide for reference. It is NOT necesary to download or create the environment.

Instructions to use the AWS environment

https://github.com/sueiras/training/blob/master/docs/aws.md

https://github.com/sueiras/training/blob/master/docs/install_tensorflow_ubuntu_aws.md

AWS AMI for this course:

• region: Ireland
• AMI id: ami-8eb287f7
• Name: sueiras-medical-images-02

Anaconda environment

Only for reference. Not necesary for the course.

1.- Install anaconda python 3.6 last version. All default options.

2.- Start the Anaconda terminal and execute...

# update package managers
conda update conda

# Create environment and install packages
conda create -n tf python=3.6

conda activate tf


conda install graphviz
conda install pandas scikit-learn
conda install -c anaconda jupyter 
conda install matplotlib
conda install pillow 

pip install Cython
pip install pydot-ng
pip install lxml

pip install --ignore-installed --upgrade tensorflow 

Download data

Blood classification dataset

Blood detection dataset

Cats & dogs dataset

Install and configure the Tensorflow object detection API on windows

Detailed instructions here

For windows users read this

Basic dependencies

• Activate the previous Anaconda environment. The API basic dependencies are already included.

Protobuf Compilation

0. Whit tensorflow 1.7 and 1.8, first solve this bug

1. Download Google Protobuf https://github.com/google/protobuf Windows v3.4.0 release “protoc-3.4.0-win32.zip” and extract

2. Move to the dir models\research.

cd <path_to_tensorflow_models>\models\research

3. Execute the protobuf compilation

"<path_to_protobuf>\protoc-3.5.1-win32\bin\protoc.exe" --python_out=. .\object_detection\protos\anchor_generator.proto .\object_detection\protos\argmax_matcher.proto .\object_detection\protos\bipartite_matcher.proto .\object_detection\protos\box_coder.proto .\object_detection\protos\box_predictor.proto .\object_detection\protos\eval.proto .\object_detection\protos\faster_rcnn.proto .\object_detection\protos\faster_rcnn_box_coder.proto .\object_detection\protos\grid_anchor_generator.proto .\object_detection\protos\hyperparams.proto .\object_detection\protos\image_resizer.proto .\object_detection\protos\input_reader.proto .\object_detection\protos\keypoint_box_coder.proto .\object_detection\protos\losses.proto .\object_detection\protos\matcher.proto .\object_detection\protos\mean_stddev_box_coder.proto .\object_detection\protos\model.proto .\object_detection\protos\multiscale_anchor_generator.proto .\object_detection\protos\optimizer.proto .\object_detection\protos\pipeline.proto .\object_detection\protos\post_processing.proto .\object_detection\protos\preprocessor.proto .\object_detection\protos\region_similarity_calculator.proto .\object_detection\protos\square_box_coder.proto .\object_detection\protos\ssd.proto .\object_detection\protos\ssd_anchor_generator.proto .\object_detection\protos\string_int_label_map.proto .\object_detection\protos\train.proto 

Add Libraries to PYTHONPATH

cd <path_to_tensorflow_models>\models\research
set PYTHONPATH=%PYTHONPATH%;C:\<full_path_to_tensorflow_models>\models\research;C:\<full_path_to_tensorflow_models>\models\research\slim

Test the Installation

cd <path_to_tensorflow_models>\models\research
python object_detection/builders/model_builder_test.py

Download a pretrained model

cd <path_to_tensorflow_models>\models\research\object_detection\models
[faster_rcnn_inception_v2 (142Mb)](http://download.tensorflow.org/models/object_detection/faster_rcnn_inception_v2_coco_2018_01_28.tar.gz)