Detection of Anemia from Conjunctival Pallor and Nail Color Using Image Processing

Reduced oxygen delivery to organs causes anemia, which is a deficiency of RBCs that are not functioning properly. Based on the World Health Organization (WHO), pregnant women are classified as anemic if their hemoglobin levels fall below 11. g/dl in the first and third trimesters, and below 10.1 g/dl in the second trimester. The ability to monitor anemia is now constrained by the intrusive nature of blood analyzers and the need for specialized training to prepare and run a sample. The model's goal is to create a smartphone-based solution for spectroscopy-based normal hemoglobin level monitoring for anemia detection. Utilizing the camera in smartphone an eye image is captured, and through imaging spectroscopy, the blood hemoglobin level can be predicted. The technique starts by identifying and removing the region of eye from the image input. Then, for locating the eye center, locate the pupil first, create and locate the contour. Conjunctival pallor, obtained by reducing the ocular image by 3.5mm from the eye center. Edge detection entails removing edges that were mistakenly detected and filtering out the noise. For extracting the region of interest (ROI), masking and image subtraction are used. The ROI's RGB values, examined using the algorithm, and the formula creates a robust forecast regarding the blood hemoglobin level, indicate the ROI is anemic or not.