Measuring Blood Glucose Concentrations in Photometric Glucometers Requiring Very Small Sample Volumes
Glucometers present an important self-monitoring tool for diabetes patients and therefore must exhibit high accu- racy as well as good usability features. Based on an invasive, photometric measurement principle that drastically reduces the volume of the blood sample needed from the patient, we present a framework that is capable of dealing with small blood samples, while maintaining the required accuracy. The framework consists of two major parts: 1) image segmentation; and 2) convergence detection. Step 1) is based on iterative mode-seeking methods to estimate the intensity value of the region of interest. We present several variations of these methods and give theoretical proofs of their convergence. Our approach is able to deal with changes in the number and position of clusters without any prior knowledge. Furthermore, we propose a method based on sparse approximation to decrease the computational load, while maintaining accuracy. Step 2) is achieved by employing temporal tracking and prediction, herewith decreasing the measurement time, and, thus, improving usability. Our framework is validated on several real data sets with different characteristics. We show that we are able to estimate the underlying glucose concentration from much smaller blood samples than is currently state-of-the- art with sufficient accuracy according to the most recent ISO standards and reduce measurement time significantly compared to state-of-the-art methods.