sHEMO: Smartphone Spectroscopy for Blood Hemoglobin Level Monitoring in Smart Anemia-Care
Sagnik Ghosal, Debanjan Das, Venkanna Udutalapally, Asoke K. Talukder, Sudip Misra
Abstract
Anemia occurs due to low blood hemoglobin levels, and for its mass screening, both invasive and smart anemia-care techniques are used. However, significant drawbacks include high costs, lack of state-of-the-art facilities, invasive techniques, and lack of smartphone implementation, using additional equipment and non-autonomous functioning, for smart anemia-care techniques. The paper proposes a novel, autonomous, smart anemia-care technique to address these problems in the form of a spectroscopy based blood hemoglobin level monitoring model. This approach uses a smartphone camera as a sensor. It quantifies the hemoglobin level based on the region of interest's color spectroscopy, which is the conjunctival pallor, extracted autonomously. Anemia is diagnosed if the predicted hemoglobin level is <; 11.5 g dL <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> . For easy smartphone implementation, the proposed model uses image analysis techniques to estimate the hemoglobin level. Our model reports an accuracy of ±0.32 g dL <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> , sensitivity of 89% compared to the actual blood hemoglobin levels (n = 65 participants). Also, the model remains robust to a wide range of illumination and the type of device used. It thereby establishes itself as a reliable and suitable replacement for the blood-based laboratory hemoglobin tests by leveraging the feature of at-home diagnosis of anemia.