Depression is by far the most common mental disorder, with more than 350 million people battling the condition everyday. This disorder is often underdiagnosed, with the average delay between the onset of symptoms and intervention being 8-10 years. Consequently, depression and anxiety disorders cost the global economy $1 trillion each year in lost productivity. Despite this, there is no easy way to measure one's mental wellness.

Three factors influence these startling statistics: the lack accessibility in accurately measuring and tracking stress/depression, the insufficiencies of integrated tools in collecting behavioral biomarkers (eg. micro facial expressions/vocal tone), and the absence of long-term tracking to monitor the effectiveness of therapies.

With 75% of mental illnesses beginning by the age of 24, how do we provide an accessible way to continuously monitor and measure an individual's stress/depression level? A review of world literature notes that depression is often linked to smaller smiles and differences in speed and pitch variation. As a result, we have produced a user-friendly AI buddy that not only tracks facial expressions but also natural language processes.

What it does

Ethera is an AI bot that is geared toward helping those with mental illneses to find the right therapy care based on facial expressions and natural language processesing.

How we built it

We designed the algorithm to combine the results received from 2 services - emotional analysis and speech analysis. The emotional analysi was conducted with the help of the google cloud vision api. The speech analysis was implemented with speech to text conversion library and stanford's core nlp models to extract features helping to get sentimental analysis. The whole services are hosted on google cloud platform.

Challenges we ran into

Some of the challenges we ran into include: the lack of understanding how machine algorithm works, figuring out which API would be appropriate for our challenge, and trying to combine the results of the two aforementioned services. In addition, we faced deployment issues to the Google Cloud Service. In terms of the business aspect, we face difficulties in ensuring HIPAA compliance and determining the source of funds for the project. Furthermore, in terms of the user-design and prototyping, we had wanted to implement animation within the application. We had decided to use a paid-GIF version to complete our project.

Accomplishments that we're proud of

Two accomplishments we are the most proud of involve learning the software engineering knowledge behind the implementation of Ethera and the deployment to Google Cluod.

What we learned

We learned how to work in an interdisciplinary team to divide the project appropriately based on our backgrounds and expertise. Because we were mainly working on different portions of the project, we ensured that we all had proper communication to deliver the results. By the end of the day, we were able to deliver a proof of concept and minimal viable product (MVP).

What's next for Ethera: the AI buddy that supports your mental well-being

Ideally, we would love to bring Ethera into fruition, creating a deployable prototype that we can first launch into less regulated markets, such as Japan, China, and Australia. These markets have a similar demographic and provide an opportunity to improve the product prior to launching in the United States.

Video Demo:

Built With

Share this project: