BRITE

Inspiration

Inspiration for our product came from having a number of skin care products that are difficult to identify just by tactile feedback. We believed technology could help to determine what products you are holding. We were also unsure of how to use our products together and were inspired to develop skin care routines that can be explained simply through audio and can recommend products that would complement our routine.

What it does

It can identify a product the user is trying to find (product identification), allows for an inventory of products that the user has previously bought/currently owns (product inventory + additional details about each product), skincare routine steps, siri navigation

How we built it

The iOS portion was developed using Swift and Xcode for the UI, logic, and Siri voice integration. The product detection feature is powered through machine learning based on a trained model from collection of product images that we annotated using a site called Roboflow. The personalized routines and recommendations are formulated with artificial intelligence via the OpenAI API.

Challenges we ran into

Adjustments were needed in the test images dataset in order to improve the accuracy to identify a product. We also realized the importance for frequent audio feedback and voice controls at each step as our target community doesn't have visual feedback.

Accomplishments that we're proud of

We're proud of developing an accessible application in a limited time frame using new data-oriented tools that we have no experience in, such as machine learning and artificial intelligence. We're proud that it not only assists the blind and visually impaired community, but also assists other groups, including us, to develop a healthy skincare routine based on our skin types.

What we learned

We were able to learn more about iOS development, how to leverage machine learning and artificial intelligence, and how to design with accessibility in mind.