Start

Capture and collect images of skin or eyes

Preprocessing & Data Augmentation

Feature Extraction (Conv + Pooling)

Classification Head (Dense + Dropout)

Check the Metrics: Accuracy, F1, ROC-AUC, PR-AUC Plot the confusion matrix

Save the final model (.keras) Streamlit the final application

Image to be tested as input

Abnormal skin or eye image is detected

Show that the person is affected by jaundice

Show that the person is healthy

YES

NO