Deep Learning and Transfer Learning for Malaria Detection

Algorithm 1

	Step 1: Importing the standardized Dataset
	Step 2: Dataset Preprocessing via Data Augmentation
	(a) Resizing
	(b) Normalization
	Step 3: Initiate Convolution Neural Networks
	% Extract features while preserving the spatial correlations of the input
	Step 4: Start CNN Model Training
	Step 5: Dataset division into two subcategories
	(c) Training set
	(d) Validation set
	Step 6: Transfer learning procedure is initiated
	% Keep all of the convolutional layers with their weights pretrained
	Step 7: Train the model in two stages
	(a) Freeze the body weights
	(b) Gradually unfreeze the layers
	(c) Fine-tuning
	Step 8: Evaluation of Model Performance
	Step 9: Identification of Malarial parasites
	Step 10: Computation and Comparison of performance metrics
	(e) Accuracy
	(f) Precision
	(g) Sensitivity
	(h) Specificity