Karlsen Haakon, Dong Tao
IEEE Trans Biomed Circuits Syst. 2017 Apr;11(2):455-463. doi: 10.1109/TBCAS.2016.2633508. Epub 2017 Mar 15.
An ambulatory pre-screening Point-of-Care (POC) device compatible with commercially available diapers has been developed to rapidly screen urine samples for incontinent or functionally impaired elderly. This POC device consists of a set of colorimetric reaction pads with accompanying reference colors. A smartphone with camera is a convenient tool for analysis of colorimetric assays; and a software application has been developed for smartphones to photograph the colorimetric assay and classify colorimetric reactions according to the reference colors. To facilitate detection of multiple biomarkers, e.g., 12 biomarkers with 2-7 references per biomarker, automatic localization of test/reference pads has been implemented through recognition of corner alignment marks and projective coordinate transformation for perspective removal. Each test run trains a set of classifiers from extracted reference data, which is used to classify the extracted test data. The smartphone application gives semi-quantitative results and functions independently of illumination intensity, illumination color, device type (smartphone brand/model), device settings (ISO, shutter speed, aperture) and automatic camera preprocessing. The smartphone application has been tested successfully on Samsung Galaxy S3, S6 Edge, S7 Edge, ZTE Nubia V7 mini and Iphone 6 in various illumination conditions.
已开发出一种与市售尿布兼容的门诊预筛查即时检测（POC）设备，用于快速筛查失禁或功能受损老年人的尿液样本。该POC设备由一组比色反应垫及相应的参考颜色组成。带摄像头的智能手机是分析比色测定的便捷工具；并且已为智能手机开发了一款软件应用程序，用于拍摄比色测定并根据参考颜色对比色反应进行分类。为便于检测多种生物标志物，例如，12种生物标志物，每种生物标志物有2 - 7个参考物，通过识别角对齐标记和用于透视消除的投影坐标变换实现了测试/参考垫的自动定位。每次测试运行都会根据提取的参考数据训练一组分类器，这些分类器用于对提取的测试数据进行分类。智能手机应用程序给出半定量结果，且其功能不受光照强度、光照颜色、设备类型（智能手机品牌/型号）、设备设置（ISO、快门速度、光圈）和相机自动预处理的影响。该智能手机应用程序已在三星Galaxy S3、S6 Edge、S7 Edge、中兴努比亚V7 mini和苹果iPhone 6上的各种光照条件下成功测试。
