Hemoglobin screening using cloud based mobile photography applications [Online First]

Abstract

Objective: This paper aims to present Selfienemia, a novel non-invasive hemoglobin estimation mobile application working under controlled illumination conditions. Materials and methods: Mobile application was supported by a parameter optimization and curve fitting backend in the cloud to provide dynamic and device independent results based on a-priori observations and the World Health Organization’s (WHO) color scale. Thirty-four and sixty-four individuals were involved during both the mobile app training and posterior blind test respectively. Results: Selfienemia (Mobile App) was found to be suitable as a non-invasive hemoglobin estimation tool. Estimation levels showed high correlation against the gold standard invasive test. Skin pigmentation and color components were found to have specific relationships labeling individuals within one Hb concentration group. Conclusion: Selfienemia may provide an effective way to screen hemoglobin in medical shortage situations. However, it still can’t replace traditional invasive testing. Future work must include testing additional fitting techniques for illumination aware scenarios as well as quantification of skin pigmentation effects on estimation results.