Peer review of researchgrant applications at the National Institutes of Health 2: review by an initial review group.

A major problem related to chronic health is patients’ “compliance” with new lifestyle changes, medical prescriptions, recommendations, or restrictions. Heart-failure and hemodialysis patients are usually placed on fluid restrictions due to their hemodynamic status. A holistic approach to managing fluid imbalance will incorporate the monitoring of salt-water intake, body-fluid retention, and fluid excretion in order to provide effective intervention at an early stage. Such an approach creates a need to develop a smart device that can monitor the drinking activities of the patient. This paper employs an empirical approach to infer the real water level in a conically shapped glass and the volume difference due to changes in water level. The method uses a low-resolution miniaturized camera to obtain images using an Arduino microcontroller. The images are processed in MATLAB. Conventional segmentation techniques (such as a Sobel filter to obtain a binary image) are applied to extract the level gradient, and an ellipsoidal fitting helps to estimate the size of the cup. The fitting (using least-squares criterion) between derived measurements in pixel and the real measurements shows a low covariance between the estimated measurement and the mean. The correlation between the estimated results to ground truth produced a variation of 3% from the mean.