Inspirations and a Beginning:
Taking root both in our own grievances and delights, Baymin is our response to the vertical of mental health based on the popular Disney character Baymax, a friendly personal care robot designed to autonomously detect the physical and emotional state of its user and provide services to better their day. As our parodic title suggests, Baymin is an app-enabled device purposed for doing the same functions as our hero at a smaller scale. It utilizes CV technology and a little extra coding to determine a user’s “mood” and offer suggestions as to how to improve their day in the short run and mental health in the long run.
In order to both scan for and interpret data input from a camera, we took advantage of the edgeIQ API from AlwaysAI, a company geared towards supporting software developers implementing machine learning, particularly in the fields of computer vision. Thanks to their innovative programming, we were able to produce coding for determining mood and physical state as well as corresponding, recommended actions. The key programs for Baymin are: object tracking, object detection, posture tracking, and pose detection. As is observable, each process comes in two parts, for a cohesive “read and respond” format.
What does it do and how?
Baymin receives inputs from a camera, scanning for objects and poses, and processes the information in through a Raspberry Pi 4. Using the edgeiq API, the data is then used as a basis for a determination of mood from a bank of basic emotions that we pre-set. Baymin also has a basic activity tracker that will encourage users to maintain healthy habits despite high-stress environments.
Who is Baymin for?
As an app-enabled device, Baymin is meant for anyone with a mobile device. Baymin is designed to be a companion who will support mental health within its user. Needless to say, Baymin requires some access to the user’s privacy on account of the data being captured vias camera, so users must realize and be able to give consent to Baymin to utilize the camera. However, with a companion always looking out for them, users will be delighted with the new company. The suggestion of no longer bearing the weight of isolation often comes as a comfort to much of the population, especially within the youth. With the vertical of mental health in mind, the ultimate goal of Baymin is to alleviate some of the ever-growing pressure imposed upon society.
While developing Baymin, we struggled mainly in the aspects of hardware. Immediately after we began working with our materials, the hardware we received began proving to be faulty, unbeknownst to us. Fortunately, we were able to replace the non-functional components and most of our only issues were based in getting accustomed to the syntax of Python, as most of us are C++ and Java coders. However, our most daunting problem came in the form of the fact that our Raspberry Pi was limited: it could not connect to the internet. As such, we could not reach the full potential of Baymin, lacking connectability to the phone with an app as well as other developments. Regardless of all this, with much collaboration and research, our team gave the best form for Baymin despite the complex constraints and restrictions.
The original intent of Baymin was to give song recommendations to users based on their determined mood, supporting their mental health with the comforting ambience of music. We ended up swapping this idea for a physical activity tracker, but we wish to continue to develop the music based origins, which can lead to partnerships with music streaming services. Additionally, Baymin would benefit from upgrades in capabilities of recognizing specific users in crowds, facial expressions, and vocal cues.