Mobile device reference apps to monitor and display biomedical information

Abstract

Smart phones and other mobile technologies can be used to collect and display physiological information from subjects in various environments – clinical or otherwise. This thesis highlights software app reference designs that allow a smart phone to receive, process, and display biomedical data. Two research projects, described below and in the thesis body, guided this development. Android Studio was chosen to develop the phone application, after exploring multiple development options (including a cross-platform development tool), because it reduced the development time and the number of required programming languages. The first project, supported by the Kansas State University Johnson Cancer Research Center (JCRC), required a mobile device software application that could determine the hemoglobin level of a blood sample based on the most prevalent color in an image acquired by a phone camera, where the image is the result of a chemical reaction between the blood sample and a reagent. To calculate the hemoglobin level, a circular region of interest is identified from within the original image using image processing, and color information from that region of interest is input to a model that provides the hemoglobin level. The algorithm to identify the region of interest is promising but needs additional development to work properly at different image resolutions. The associated model also needs additional work, as described in the text. The second project, in collaboration with Heartspring, Wichita, KS, required a mobile application to display information from a sensor bed used to gather nighttime physiological data from severely disabled autistic children. In this case, a local data server broadcasts these data over a wireless network. The phone application gathers information about the bed over this wireless network and displays these data in user-friendly manner. This approach works well when sending basic information but experiences challenges when sending images. Future work for both project applications includes error handling and user interface improvements. For the JCRC application, a better way to account for image resolution changes needs to be developed, in addition to a means to determine whether the region of interest is valid. For the Heartspring application, future work should include improving image transmissions.