A new method and a non-invasive device to estimate anaemia based on digital images of the conjunctiva

Medical practice for determining haemoglobin (Hb) concentration, which is important for anaemic patients in need of blood transfusion, requires a blood sample. In this paper, we propose a non-invasive approach to Hb estimation based on the image analysis of a specific conjunctival region. Our goal is to develop a device that is not expensive and simple to use for assessing the anaemic condition; this device could be used by the physician to decide whether to take a blood sample or even by a patient at home to decide whether to inform a physician; in this way, we can avoid having the patient go frequently to the laboratory to take a blood sample. This device also allows us to rapidly screen for anaemia in a large number of persons, for example, groups of girls who are at the beginning of the menses. Here, we detail the prototype of our device and the methodology for extracting key information from the colour values of the acquired image. Tests conducted on 113 anaemic and healthy persons show a strong correlation between the real Hb value obtained through blood sampling and the value estimated by our device. A k-nearest neighbor (kNN) classification algorithm for assessing the (non)anaemic condition yielded good results and helps doctors avoid a significant number of blood tests.