Course Syllabus

Welcome to BIOMEDIN 260/RAD 260!

 

Overview:

Biomedical imaging is an exploding field. The technologies for visualizing the body (the imaging modalities) are becoming very powerful, providing exquisite images of tissue morphology, revealing tissue function, and even beginning to see molecular events such as gene expression. Imaging is at the core of medical practice; nearly all patients have imaging of some sort during care, and many studies produce thousands of images. Just as the genetic data explosion has fueled the field of bioinformatics, the growth in digital imaging is necessitating techniques in imaging informatics.

Imaging Informatics is the science of analytic, storage, retrieval, and interpretive methods to optimally use the increasingly voluminous imaging data in biomedicine, and integrate and understand them in the context of complementary molecular and clinical information to improve clinical diagnosis and therapy in medicine. Imaging informatics spans a broad spectrum of topics that include engineering, computer science, statistics, radiology, and medicine. This course will provide a broad overview this field as well as the foundation techniques required to process, analyze, and use images for scientific discovery and applications.

Topics covered in this course: 

  • Types of imaging methods and how they are used in biomedicine
  • Image processing, enhancement, and visualization
  • Computer-assisted detection, diagnosis, and decision support
  • Access and utility of publicly available image data sources
  • Linking imaging data to clinical data and phenotypes
  • Computer reasoning with images
  • New questions in biomedicine using imaging informatics. Case studies.

 

Detail:

Also known as: RAD 260 

Units: 4 (or 3 with consent of instructor) 

Lectures: Spring Quarter 2018-2019, Monday and Wednesday 1:30 PM - 2:50 PM, Gates Building, Room B3 

TA-led Section: MSOB X-393 (3rd floor), 10:30 AM - 11:30 AM, Fridays, starting 4/5/2019

First class: Monday, April 1, 2019 (no, this isn't an April Fool's joke) 

Prerequisites: Programming ability at the level of CS 106A, familiarity with statistics, basic biology, knowledge of Python (highly recommended), or approval of the instructor. 

Grading: Assignments (45% total), Participation (10%), Midterm (15%), Final Project (30%) 

Participation: There are many different ways to participate, including but not limited to attending class, attending section, asking questions, and contributing to discussions on Piazza.

 

Staff:

Instructors:

Part I: David Paik, PhD, Adjunct Lecturer, Radiology
david [dot] paik [at] stanford [dot] edu
Part II: Daniel Rubin, MD, MS, Associate Professor Profile
dlrubin [at] stanford [dot] edu
Teaching Assistants:
Kevin Thomas (kathoma [at] stanford [dot] edu)
Guhan Venkataraman (guhan [at] stanford [dot] edu)

 

Office Hours:

Bring questions to TA-led sections.
Location: MSOB X-393 (3rd floor)
Time: 10:30 AM - 11:30 AM, Fridays, starting 4/5/2019
Professor Rubin: by appointment (email Kimberly Wilderman (kimwild [at] stanford [dot] edu) to set up a meeting time). MSOB X-335.
Professor Paik: by appointment (email Professor Paik to set up a meeting time).

Contact/Questions: Most questions should be posted to our Piazza page so that all students can benefit from the answers. Other queries can either be emailed privately to TAs or Professors.

 

Audience:

This course is designed for:

• Graduate students in biomedical informatics, computer science, bioengineering, or other related disciplines
• Medical students
• Medical, pediatric, surgical or other fellows with an interest in learning and using imaging informatics in research
• Interested undergraduates
• Auditors welcome including medical staff, medical/pediatric/surgical fellows, post-doctoral fellows, and undergraduates. Please contact us to be added to the class email list.

 

Lectures:

The video lectures can be accessed via SCPD.
All handouts, assignments, etc. will be released on Canvas.
Date Topic Lecturer Lecture Notes Readings
04/01/2019 Course Intro Daniel Rubin 01_Introduction.pdf Biomedical Imaging Modalities
04/03/2019 Visualization David Paik 02_Visualization.pdf Marching Cubes
04/08/2019 Convolution and Filtering David Paik 3 Segmentation.pdf

Canny Edge Detection

Anisotropic Diffusion

04/10/2019 Image Segmentation David Paik 4 Filtering.pdf Level Sets
04/15/2019 Computational Feature Extraction David Paik 5 Geometric Features.pdf
04/17/2019 Image Features David Paik 6 Texture Analysis.pdf

Haralick Features for Image Classification

Saito Exact Distance Transform

04/22/2019 Image Registration David Paik 7 Registration.pdf

Image Registration Review

Rotating Objects Using Quaternions

Homogeneous Coordinates

Mutual Information Registration Review

Analysis of Vasculature for Liver Surgical Planning

Talairach Atlas Labels for Functional Brain Mapping

04/24/2019 Classical ML for Images David Paik 8 Radiomics.pdf
04/29/2019 DL for Images David Paik 9 DL for images
05/01/2019 Evaluation of Machine Learning Francisco Gimenez
05/06/2019 MIDTERM
05/08/2019 Data Integration  Mirabella Rusu
05/13/2019 Semantic Feature Extraction: An Intro Daniel Rubin Semantic Feature Extraction
05/15/2019 Semantic Feature Extraction: NLP Daniel Rubin NLP Semantic Feature Extraction
05/20/2019 DL for NLP in Radiology Imon Banerjee Code on Github
05/22/2019 Neuroimaging Killian Pohl
05/27/2019 MEMORIAL DAY
05/29/2019 Medical Imaging Applications of DL Daniel Rubin 16-Medical Imaging Applications of DL.pdf
06/03/2019 Computer Reasoning and Decision Support Daniel Rubin 17 Computer_Reasoning_Images.pdf
06/05/2019 DL in Radiology Matt Lungren

 

Textbooks:

Required books: None (the course will be taught using recent publications) 

Recommended books:

Digital Image Processing in Matlab - Gonzales and Woods. Also see here. 
Insight into Images - Yoo 
Digital Image Communications in Medicine - Pianykh 
Practical Imaging Informatics - Branstetter 
Naked to the Bone - Bettyann Kevles 

Useful datasets

 Image Datasets

 

Coursework:

Description Out Date Due Date Percent
Problem Set 1 Lung Field Segmentation 04/05/2019 04/19/2019 ~15%
Problem Set 2 Feature Extraction from Mammograms 04/15/2019 05/10/2019 ~15%
Problem Set 3 Machine Learning with Mammography 05/06/2019 05/24/2019 ~15%
Final Project Proposal | Milestone | Presentations | Writeup 04/05/2019

Proposal: 04/26/2019

Milestone: 05/17/2019

Presentations: 06/10/2019

Writeup: 06/10/2019 

30%
In-Class Midterm 1:30-2:50 PM, Gates B03 - 05/06/2019 15%
Participation In class, during OH, and/or on Piazza - - 10%

Schedule

A detailed schedule for the quarter (including lecture topics, relevant dates, guest lectures, section topics, and Finals week details) can be found here.

Course assignments

During the quarter, the students will undertake developing and implementing four substantial imaging informatics applications (3 problem sets and 1 project), increasing in difficulty during the quarter. Students will turn in problem sets as *.pdf conversions of their Jupyter notebooks. Grades for these assignments will be released on Canvas and weighted as in the table above; midterms and finals may be graded on a curve.

Due dates and Late Day Policy

Assignments are due at or before 11:59 PM on the dates indicated avove, and new assignments are released on the Monday before the old assignments are due. You may have up to 4 late days in the quarter that you can use only on Problem Sets and not Final Project assignments. No extra credit will be given for unused late days.

Once you are out of late days, you will lose 10% of your grade on late Problem Sets each day past the deadline (each "day" rounds up - so if you are one hour past the deadline, that is a day).

Final Project assignments turned in late will automatically lose 10% of the grade each day past the deadline.

Please contact the TAs for special accommodation if applicable.

Midterm

Based on the lectures and the readings. Open book. To be held during class, Monday, May 6th, 2019.

Final Project Presentations

To be held on Monday, June 10th, 2019 from 3:30 PM - 6:30 PM (Location TBD).

Questions

Students are encouraged to post questions on Piazza.

Other Considerations

This course will use Python 3 for all programming-related assignments. We recommend using the Anaconda distribution as it comes with many useful scientific computing packages. It also allows you to use Jupyter notebooks, which will be used for submitting assignments.

`ssh` and `cardinal`

You will need to have access to email (be sure you're registered on Axess so that you get email announcements sent to the course list), the course website, and the Stanford cardinal machines. All of these resources are available to Stanford students at Sweet Hall and elsewhere as well as through remote (`ssh`) access. To log in, you will need to use your SUNet ID. If you don't have a SUNet ID, get one here ASAP. To log in to the "cardinal" cluster machines, use a secure shell (ssh). Windows: You will have to download a terminal emulation that allows ssh. Stanford offers a few free ones here; a popular one is PuTTY. Directions for using Putty to connect to cardinal:
  1. Under "Host Name", enter cardinal.stanford.edu
  2. Under "Protocol", choose SSH
  3. Press the "open" button.
A terminal window should appear, connected to cardinal. Putty will tell you if there was an error.
On OS X, Unix, Linux:
  1. Open a terminal window
  2. Type "ssh sunetid@cardinal.stanford.edu"
For more information on the various campus computers you can access:

Collaboration policy

You may work on the three problem sets with up to one other person and indicate as such when turning the assignments in. Midterms are individual efforts. You can talk with others in the class about the class project, but your group must turn in its own individual work.

Honor code

Stanford University's Honor Code will be observed and uphold at all times.

Students with Documented Disabilities

Students who may need an academic accommodation based on the impact of a disability must initiate the request with the Office of Accessible Education (OAE). Professional staff will evaluate the request with required documentation, recommend reasonable accommodations, and prepare an Accommodation Letter for faculty dated in the current quarter in which the request is made. Students should contact the OAE as soon as possible since timely notice is needed to coordinate accommodations. The OAE is located at 563 Salvatierra Walk (phone: (650)-723-1066).

Course Summary:

Date Details Due