Commercial Platforms for Healthcare Analytics: Health Issues for Patients with Sickle Cells
J.K. Adedeji, Taoreed Olakunle Owolabi, Rufus Sola Fayose · 2022
This research has applied the principles of convolutional deep learning algorithm and image classification to solving problems relating to sickle cell disease, causing cardiovascular disorder in patients living with it. The demise due to ignorance of many who lives with this disorder in developing countries in recent is on the high side. Hence, there is the need to have an advanced computerized biological image processing device that can provide effective information and assist in monitoring the health status of SCD patients in our communities. In order to achieve this, the CNN classifier algorithm with fully connected layers coded in Python on a PyTorch platform was used on 300 images. The pictures of the red blood cells were taken with electron microscope from 60 normal people and re-sized up to 224 × 224 dpi in resolution. The training epoch was 10 to ensure non-overfitting and pre-trained using the ResNet algorithm. The values of the training and validation sets after adjustment are 0.020 accuracy error and the dropout at 0.022036 best value accuracy. The chapter concluded that the image classifier is suitable for monitoring the health status of sickle cell anemia patients and processing of any complex biological images to obtain information relevant to any abnormality especially patients in sickle cell crisis.