Deep Learning Regression-Based Retinal Layer Segmentation Process for Early Diagnosis of Retinal Anamolies and Secure Data Transmission through ThingSpeak

Algorithm 1

Proposed regression algorithm.

	Step 1: Import retinal OCT and fundus images from the database
	Step 2: Preprocess the images by applying Gaussian and bilateral filters
	%Gaussian filter as below equation for smooth impulse response and no ringing effects
	Step 3: Image segmentation via Chan–Vese procedures
	Step 4: Object clustering via K- means clustering(i)Choose “k” cluster centres at random.(ii)Determine the distance between each data point and the cluster centres.(iii)Assign the data point to the cluster centre with the shortest distance from the cluster centre among all cluster centres.(iv)Calculate the new cluster centre again.
	Step 5: Edge detection and boundary tracking to identify lesions in OCT and fundus images.
	Step 6: Carry out morphological enhancement for filling the gaps in objects
	Step 7: Feature extraction and classification using regression deep-learning networks
	Step 8: Transmission to remote areas via the IoT module