AI for Medical_1_Diagnosis

This repository contains the notes, codes, assignments, quizzes and other additional materials about the course "AI for Medical Diagnosis" from DeepLearning.AI Coursera. Enjoy!

The notes contain the modules outlined below:

Week	Module	Gist
1-1	AI for medical specialization	Demo Pre-requisites
1-2	CV to medical diagnosis	Medical iamge diagnosis Eye disease and cancer Lab-Data exploration and image pre-processing
1-3	Class imbalance and small training sets	Model building Training, prediction, and loss Image classification and class imbalance Binary cross entropy loss function Impact of class imbalance on loss calculation Lab-Counting labels and weighted loss function Resampling to achieve balanced classes Multi-task loss, dataset size, and CNN architectures Lab-Densenet A small training set Generating more samples
1-4	Check model performance	Model testing Splitting data by patient Lab-Patient overlap and data leakage Sampleing Ground Truth and consensus voting Additional medical testing
2-1	Key evaluation metrics	Sensitivity, specificity and evaluation metrics Accuracy in terms of conditional probability Sensitivity, specificity and prevalence PPV, NPV Confusion matrix
2-2	Threshold and Evaluation Metrics	ROC Curve and Threshold Varying the Threshold Lab-ROC Curve and Threshold
2-3	lnterpreting Confidence lntervals Correctly	Sampling from the Total Population Confidence Intervals 95% Confidence Interval
3-1	MRI data	Iamge segmentation Lab-MRI data and labels
3-2	Image segmentation	Image registration and segmentation Extract a sub section Convolutional Neural Network 2D U-Net and 3D-Unet Data Augmentation for segmentation Loss function for image segmentation