IFlick: Smartphone-based anemia screening in rural healthcare paradigm

Abstract

Anemia is the most prominent and severe disease, the prevalence of low red blood cells. Identifying and continuously monitoring hemoglobin in the human body is a challenge. Anemia detection involves various invasive and non-invasive solutions. However, both solutions have limitations such as cost, time-consuming, resource dependencies, and not being implemented in smartphones. Therefore, anemia identification with limited resources, low cost, more reliable, and edge intelligence is required. This paper proposes a non-invasive and edge-based point-of-care solution (iFlick). The solution estimates hemoglobin level by performing automatic region of interest segmentation and histogram calculation of RGB colors from the eye images. While the solution is tested in the primary health centers, a person is anemic if the hemoglobin level ranges < 11.5 gdL −1 . The proposed application reported an accuracy of ± 0 . 057 gdL −1 with 94% sensitivity for 157 participants. The experiment found that iFlick takes 10 - 15 s to estimate the hemoglobin value. The proposed solution helps the rural healthcare system determine hemoglobin values with the iFlick smart mobile application.