Program*

*preliminary version, this schedule is subject to slight modifications

 

Monday June 29, 2020

08h30 - 09h30

Registration

(Christian Desrosiers)

09h30 - 09h45 Welcome talk
09h45 - 12h00

Introduction to machine learning (Training, testing, validation, metrics, over/underfitting, linear vs non-linear classification, applications in medical imaging, ...)

(Jose Dolz - ETS)

12h00 - 13h30 Lunch (at the ETS)
13h30 - 15h00

Basics of deep learning Part 1(Perceptron, stochastic gradient descent, learning rate, logistic regression, cross-entropy, multy-layer perceptron, activation functions, multiclass perceptron)

(Pierre-Marc Jodoin - Université de Sherbrooke)

15h00 - 15h30 Break
15h30 - 17h00

Basics of deep learning Part 2 (Forward and backward propagation, batch size, Convolution neural nets (CNN), feature maps, pooling, applications.)

(Pierre-Marc Jodoin - Université de Sherbrooke)

17h30 - 19h00 Cocktail at the ETS


Tuesday June 30, 2020

09h00 - 10h30

Advanced concepts in deep learning Part 1(CNN architectures for classification - AlexNet, VggNet, ResNet, InceptionNet, ..., Localization networks  - FasterRCNN,Yolo,SSD- Segmentation methods-EncoderDecoder, UNet, ENet,...)

(Christian Desrosiers - ETS)

10h30 - 11h00 Break
11h00 - 12h30

Advanced concepts in deep learning Part 2 (Recurrent neural networks, visualisation, explainability)

(Christian Desrosiers - ETS)

12h30 - 14h00 Lunch (at the ETS)
14h00 - 17h00

Hands-on session: classification of medical images

(Olivier Bernard)

(Thomas Grenier)

(Pierre-Marc Jodoin)

Format: python/ jupyter notebook

Data:2D images from MRIs and CT

Content: introduction to flyodHub (GPU cloud) and keras/tensorflow.

 

Wednesday July 1st 2020

09h00 - 17h00 Holyday.  ETS closed... Have fun visiting the town!

 

Thursday July 2nd, 2020

09h00 - 10h30

Generative and adversarial methods for medical imaging part 1 (GANs and autoencoders.)

(Mohammad Havaei - Imagia inc.)

10h30 - 11h00 Break
11h00 - 12h30

Generative and adversarial methods for medical imaging part 2 (Domain adaptation, adversarial learning, adversarial attacks )

(Mohammad Havaei - Imagia inc.)

12h30 - 14h00 Lunch (at the ETS)
14h00 - 17h00

Hands-on session: Automatic segmentation of 2D echocardiographic images by deep learning

(Olivier Bernard)

(Thomas Grenier)

(Pierre-Marc Jodoin)

Format: python/ Jupyter Notebook

Data: database from 500 patients of 2D sequences (from two different orientations) named CAMUS database. This database is public and we will be able to use it (and promote it) during the school.

Content: implement step by step a convolutional neural network (U-net) to automatically segment 2D echocardiographic sequences (both 2 chamber and 4 chamber view orientations).

17h00 - 18h30*

Group visit of the Musée Pointe-à-Callière on the history of Montreal.

18h30 - 21h00* Dinner cocktail at the Musée Pointe-à-Callière.

* Evening schedule subject to slight modifications.

 

 

Friday July 3rd, 2020

09h00 - 10h30

Graph CNN

(Herve Lombaert - ETS)

10h30 - 11h00 Break
11h00 - 12h30

Weakly supervised learning

(Ismail Ben Ayed - ETS)

12h30 - 14h00 Lunch (at the ETS)
14h00 - 17h00

Hands-on session:  Weakly supervised segmentation

(Olivier Bernard)

(Thomas Grenier)

(Pierre-Marc Jodoin)

Format: python/ Jupyter Notebook

Data: MRI cardiac short axis images (ACDC 2017 dataset)

Content: Segment MRI images based on weakly annotated groundtruth.