Detect Disease using Deep Learning and Transfer Learning(ResNet50)
COVID-19 Classifier from X-Ray Images
Tasks
[✔️] Exploratory Data Analysis
[✔️] Image augumentation
[✔️] Base CNN model accuracy calculation
[✔️] Base CNN model with lower imbalance data
[✔️] RESNET 50 model accuracy calculation
[✔️] EfficientNet B4 accuracy calculation
[✔️] AUC Score comparision
[⚫] Results
Version Information v1 :
Completed exploratory data analysis of given metadata Completed exploratory data analysis of provided images Inferences of both EDA explained
v2 :
Code cleaning Output cleaning
v3 :
Completed Image Augumentation using Keras ImageDataGenerator Completed training of base CNN model on data Accuracy inference of base CNN model completed
v4 :
Trained base CNN model on balanced data Inferenced accuracy of base CNN model on balanced data Trained ResNet 50 model on data Inferenced accuracy of ResNet 50 model
v5 :
Because of severe class imbalance, metric for model training and validation is changed from accuracy -> AUC ROC Really Good Article on choosing evaluation metrics Trained EfficientNet B4 model on data Inferenced accuracy of EfficientNet B4 model AUC score comparisions of all trained models
v6 :
Added multiple metrics for better view of model comparison Added numpy and tensorflow seeding for reproducible results Code cleaning, debugging