Illumination and device independence for colorimetric detection of urinary biomarkers with smartphone.

Diaper wearing elderly with functional impairments and/or incontinence is at high risk of contracting urinary tract infections. Nurses struggle with collection of urine samples for analysis. Therefore, a smartphone application is under development as a rapid screening device to work in conjunction with a colorimetric diaper assay that collects and tests urine within a diaper. The focus of this work is to make a practical and useful tool that medical personnel (the user) can use for rapid screening on patients in the field, only with a smartphone and assay available. The main challenge is to achieve illumination and device independency for a wide range of colorimetric biomarkers without using any standardization, such as attachments, special lamps, or boxes. We achieved illumination and device independent semi-quantitative detection by using discriminant analysis and classification of simultaneously photographed colorimetric test results and reference colors to ensure that any variation in image conditions applies approximately equal for reference and test. This requires retraining of classifiers on each analysis, but appears to be the most viable solution to solving the challenges while maintaining the user-friendliness.