UAH - Americas Datafest

SUMMARY
Health24 is a healthcare application tailored for migrants. It is available for the Web, iOS, and Android platforms, and it has three major features:

Identification of medicine needed based on symptoms
Drug database
Image-text translation of medicine labels and instructions

The primary goal of Health24 may be described as follows: "Help migrants receive basic diagnoses and correctly identify and use over-the-counter (OTC) drugs to treat symptoms."

FULL DESCRIPTION
The Health24 team determined the functional pieces for a typical use case. A user is feeling sick and needs to determine what medicines to take. For that they use the symptom checker. Secondly, after knowing what medicine to take, the user would want to view the information about the drug in their native language to see how to take, what not to mix it with and so on. After they know info about the drug to take, they would go to buy the medicine. They can then take pictures of the medicine information and have the image converted to text and translated to their native language all in one step.

Before starting development we setup a BitBucket Git repository for all of our team code. We also setup a Debian Linux server to house the project, installing apache, configuring the firewall, and linking it to our repository. We installed the MySQL database for storing all of the site metrics. We then installed all of the software packages needed for the project.

We initially focused most of our work on the optical character recognition and language translation, as that was the most technically challenging. Since neither of these components existed together, phantomJS was chosen as a middle layer for combining the two components. A phantomJS script was created that leveraged a Tesseract OCR Engine and Google Translate. The script was able to process an image in English and translate the text to a target language. Then, we first focused on the web interface, as that would make this capability available to the widest audience. We accomplished this task by having a php page call the PhantomJS script in the background.

After creating the web interface we wanted to support different devices, so we created a network service (in Java) that any device could connect to in order to send a picture and the desired output language. Then, we created apps for the iOS and Android.

Simultaneously to the development of the mobile apps, one of the developers began working on the about page while another worked on the database and the metrics visualizations, creating the dynamic charts for the metrics that will be tracked by the service. We decided we did not have time to complete the first 2 steps before the deadline, so we instead stubbed in step 1 as more of a proof-of-concept. We embedded a Google Translate Javascript to the site enabling translating every page to any language, though in the future we plan to have the pages automatically translated to the user's selected language without further necessary actions.

Future Plans