Non-Invasive Anemia Detection System Using Eye Palpebral Conjunctiva Image With Raspberry Pi

The deficiency of hemoglobin or RBC cells is characterized as Anemia, which is a common medical condition around the world. Early detection and management of anemia is essential for both effective therapy and avoiding complications. In this work, we propose estimation of anemia from eye conjunctival image through Raspberry Pi. The conjunctiva can be used for non-invasive anemia testing since it is an easily accessible tissue that represents systemic hemoglobin levels. A collection dataset of ocular Palpebral conjunctiva photos from both healthy people and the hemoglobin deficiency people with various levels of anemia was gathered through camp. Our test results showed that the CNN model could accurately estimate anemia from conjunctiva images with an accuracy of about 85%. Then it predicts heamoglobin value with 99.7% accuracy. Our results imply that this deep learning model with Raspberry Pi has potential for automated, non-invasive anemia estimate, with the hemoglobin value of person's eye image which provides a useful tool for medical practitioners in contexts with low resources when conventional diagnostic techniques are intrusive or might not be feasible.