I. Course Information

cs8395_clean Deep Learning is prevalent since 2015 in medical image computing. This class covers the theories and practices of Deep Learning related to medical image computing. Instructor: Yuankai Huo, Ph.D. Guest Lecturer: Vishwesh Nath Class Meets: Tuesday & Thursday, 4:00 pm - 5:15 pm, FGH 211 Office Hours: Tuesday & Thursday, 3:00 pm - 4:00 pm, FGH 371 (or nearby tables) Contact Instructor: yuankai.huo@vanderbilt.edu Course Website: https://my.vanderbilt.edu/cs8395fall2018 Submission & Discussion: https://www.vanderbilt.edu/brightspace

II. Schedule




Aug 23 Overview of Deep Learning in Medical Image Computing slides
Aug 28 Neural Networks and CNN slides, reading1
Aug 30 Classification (Medical Image Diagnosis) slides, reading2
Sep 04 Detection (Landmark Localization and Detection) slides, reading3
Sep 06 Segmentation (Medical Image Segmentation) slides, reading4
Sep 11 GAN (Medical Image Synthesis) slides, reading5
Sep 13 Assignment 1 Presentation
Sep 18 Guest Lecture: Run Pytorch and ACCRE GPU slides
Sep 20 Guest Lecture: Dual Networks slides
Sep 25 Multi-modal Learning slides, reading6
Sep 27 Multi-task Learning slides, reading7
Oct 02 Beyond 2D: 3D Networks slides, reading8
Oct 04 Semi-/weakly-supervised Learning slides, reading9
Oct 09 Assignment 2 Presentation
Oct 11 Unsupervised Learning slides, reading10
Oct 16 Mid Term Exam exam
Oct 18 No Class: Fall Break!
Oct 23 Final Project Proposal example
Oct 25 Data Augmentation and Preprocessing slides, reading11
Oct 30 Attention Mechanism and Postprocessing slides, reading12
Nov 01 Spatial-temporal Model: RNN and LSTM slides, reading13
Nov 06 Image Retrieval and Active Learning slides, reading14
Nov 08 Assignment 3 Presentation
Nov 13 Co-learn from Medical Image and Clinical Data slides, reading15
Nov 15 Online Learning slides
Nov 20 No Class: Thanks Giving!
Nov 22 No Class: Thanks Giving!
Nov 27 Summary
Nov 29 Final Project Presentation
Dec 04 Final Project Presentation
Dec 06 Final Project Presentation

III. Assignments



Due Date

Reading Assignments Template Begining of the class
Assignment 0: Eligibility Test PDF Aug 28 2018, 4:00 pm
Assignment 1: Detection PDF Sep 13 2018, 9:00 am
Assignment 2: Classification PDF Oct 07 2018, 9:00 am
Assignment 3: Segmentation PDF Nov 15 2018, 9:00 am
* Teamwork is not allowed for assignments. An example of data_loader.py in PyTorch is provided.

IV. Assignments, Mid Term Exam, and Final Project

More details are provided Here.

V. Computational Resource

GPU computing is required for this class. I strongly recommend to use your own/lab's GPU since that is the most convenient way of writing and testing code with GUI. However, if you don't have any GPU computational resource, you can use the GPU on ACCRE by filling the following two forms: 1. ACCRE registration form (Please choose "CS839502" group) 2. GPU requirement form If you are already using ACCRE, you can submit a helpdesk ticket instead to request access to your class group. There is a relatively large demand for GPU resources at ACCRE, so limited GPU resources will be provided for this class. *The deadline of ACCRE resource application is the due date of assignment 1.


1. The class is full. Can I still get in? It is unlikely except other students drop it during the first week. 2. What is pre-requirement? Linear algebra, programming in python, introduction in machine learning. 3. Can I sit in class without registering? Yes after getting the instructor's approval. Another option is to register to audit the class (just $50).

VII. References

* We used images and contents in the slides from the following resources, thanks for the great work done by the smart people! http://deeplearning.cs.cmu.edu/ https://www.cs.princeton.edu/courses/archive/spring16/cos495/ http://ttic.uchicago.edu/~shubhendu/Pages/CMSC35246.html https://www.cc.gatech.edu/classes/AY2018/cs7643_fall https://www.deeplearningbook.org/lecture_slides.html http://introtodeeplearning.com/