M.Tech. (Online)

DS 261o Artificial Intelligence for Medical Image Analysis 3:1 (January 2023)

Course Instructor: Phaneendra Yalavarthy, CDS

Learning Outcomes:

On successful completion of the course, the student should be able to:

Identify the basic concepts, terminology, theories, models and methods for medical image analysis using artificial intelligence
Characterize the unique challenges associated with various types of medical image data modalities
Describe and implement the commonly used artificial intelligence methods used for medical image analysis.
Develop and systematically test a number of methods for medical image analysis using artificial intelligence methods
Choose appropriate evaluation methodologies to evaluate the performance of artificial intelligence methods on biomedical image analysis problems
Identify limitations of the methods covered in the course

in order to:

Curate medical image data for use in artificial intelligence based methods
Implement, analyze and evaluate biomedical image analysis systems using artificial intelligence
Use the knowledge acquired in the course read and profit by literature in the area

Topics

Overview of biological and medical imaging modalities and research/clinical applications
OverviewQuick introduction to: (a) medical imaging tools (viewers, formats, etc.) and (b) Pytorch/Tensorflow
OverviewBasic Mathematics (Linear Algebra, Probability, and Optimization)
OverviewChallenges in biomedical image data handling and curation
OverviewDetection / Segmentation / Image classification for biomedical images
OverviewMachine Learning Methods: SVM, PCA, KNN, FCM, etc. applied to medical image analysis
OverviewOverview of Neural networks and Deep Learning: Principle of learning, CNNs, Loss Functions, etc. for medical image analysis
OverviewTransfer learning, fine tuning, and generalization in medical image analysis
OverviewEvaluation methodology in medical image analysis: metrics, calibration, uncertainty, bias, etc.
OverviewChallenges in healthcare AI deployment: reproducibility, interpretability and regulatory
OverviewUnsupervised/self-supervised learning in biomedical image analysis
OverviewGenerative models and Inverse problems in medical imaging

Laboratory Component:

Systematically test a number of methods for medical image analysis using artificial intelligence methods
Mini Project to solve a test problem in medical image analysis (Ex:- Segmentation of Hepatic vessels in kidney using X-ray CT images)

Textbooks / References

Probabilistic Machine Learning: An Introduction (Adaptive Computation and Machine Learning series), Kevin P. Murphy, The MIT Press, March 2022.
Deep Learning Illustrated: A Visual, Interactive Guide to Artificial Intelligence, Jon Krohn, Grant Beyleveld, and Aglae Bassens – Addison Wesley, 2019.

Prerequisites: Basic knowledge of Systems and Signals, Proficiency in Python.

Assessment: