Head Pain Diagnostics

Diagnostics tree of the program
Main logic of the program
Diagnostics tree: different keywords trigger different parts of the program, as well as proceeding to a different part of the diagnosis
output of the program
hospital array. fetches whether a hospital is nearby given a zip code

Inspiration

Sports can have great benefits, but they often also pose significant personal risks to participants. With Optum’s goal of helping people live healthier lives in mind, an innovative product will be created to tackle ways of improve sports related healthcare. Considering the context of MinneHack and its proximity to Super Bowl LII, projects are encouraged to focus on football specifically

What it does

The Head Trauma Speech Based Health Assistant analyzes input data (from the user) and generates a diagnosis on the user’s condition. Later, if the user decides to provide the zip code, nearby hospitals are displayed. Ultimately, the goal was that it provides clear treatment options and helps the user better understand their condition.

How we built it

We used many python libraries to speed up the process, some in the stadard and some not. A list would json, pyaudio, pyhon-text-to-speech x3, PocketSphinx, and SoundRecognition.

The functions resultsCheckOR/ resultsCheckAND search for a certain word or phrase in user voice input,which triggers other functions to start (See diagnostic tree diagram)

Series of nested If / Else statements that represent tree method.(see main logic diagram) Using the “printv(tts, …)”, we were able to print the text and have the assistant speak to the user. Different statements call functions that report to the user the information and other additions based on the diagnosis.

Lastly, the program sorts and returns the top 3 possibilities of illness based on diagnosis. Using a government data collection, the position of hospitals were used to provide the user addresses to the nearest medical centers

Challenges we ran into

It is certain that the use of a faulty API prevented us from demoing this project correctly. For example, many times the API itself misinterpreted inputs from the user. Other times, the API would cease to work, requiring a full reinstallation. Further research on a more reliable speech to speech API is required.

Accomplishments that we're proud of

We are proud that we implemented a working diagnosis model related to in home health care. We were also happy to reverse engineer zip code location to local hospitals within a given area.

What we learned

We were able to learn aboyt third-party python libraries for text-to-speech and speech-to-text. We were also able to discover ways of reverse engineering a goal given something that is known.

What's next for Head Trauma Diagnostics

AI and Assistants for sport-related injuries will have a significant impact on ensuring athletes are receiving the efficient diagnosis they deserve. Obtaining an early and accurate diagnosis can lead to several benefits in the recovery process, which would save the money of several athletes and organizations.