My father suffered a stroke. He survived but had partial facial paralysis. He needed to relearn how to control his facial expressions.
What it does
CheeseFace is an electronic physical therapist. It collects biofeedback data, evaluates the quality of a smile, tracks the efficacy of the treatment, makes suggestions on a treatment plan.
How I built it
Invision, Trueface.ai, Kaggle, Google Cloud Platform, Microsoft Cognitive Services, and Microsoft Azure Machine Learning Studio.
Challenges I ran into
Locating a source of facial images. Massaging the data into a format acceptable to Azure Machine Learning Studio.
Accomplishments that I'm proud of
Awesome teamwork exhibited by each teammate.
What I learned
We learned new publicly available resources (ImageNet, Kaggle, etc.) and how to tap into them.
What's next for cheeseface
Our MVP will require a smile engine that scores the authenticity of a smile. It must take into account age, sex, personality and other variables. Potential applications are in health services, security, vetting of strangers for social interactions (dating) and business relationships.