Detection of Anemia from Palpebral Image of Anterior Conjunctiva Using SVM Classifier

Anemia is a medical disorder that arises when an individual's blood is deficient of sufficient mature, developed red blood cells with normal hemoglobin level. One of the main components of erythrocytes, hemoglobin, has a strong affinity for binding oxygen, which is necessary for cell survival. The body's cells won't get enough oxygen if there are any aberrant red blood cells or low hemoglobin levels, and this condition eventually leads to a condition known as hypoxia. The conventional method of diagnosing anemia involves pricking of finger and using analytical reagents to investigate is quite time consuming and requires skilled laboratory procedures and personnel. To overcome these limitations, a novel approach for the automated non-invasive detection of anemia is developed. In this work, a method for diagnosing anemia by analyzing changes in anemic individuals' anterior conjunctival pallor based on processed eye images is proposed. Nowadays, patients with anemia disease present in the world increased by around 60-70% respectively. This proposed work has successfully characterized to introduce novel approach for early and accurate anemia disease diagnosis. It employs LBP texture analysis for classification of eyelid images. There are several features which are considered based on extracted statistical analysis. The classification results demonstrate that these features are utilized to identify normal and abnormal patients successfully with an accuracy rate of 91%.