Pointy

Inspiration

We would like to help blind people navigate the world like we do

What it does

Helps people recognize objects and text which may have influence in their daily lives.

How we built it

We have used various machine learning libraries from python and have used a Dragonboard 410c to optimize our user experience.

Challenges we ran into

A complete new system full of challenges which had to be solved before actually implementing the system.

Accomplishments that we're proud of

Despite of it being a prototype and therefore not perfect, the product in it's entirely is a success in terms of the goal it aims to accomplish. The whole pipeline (from raw images to speech) is functional.

What we learned

We have tackled a new IoT problem, which was completely new experience for all of our team members to solve. Besides from learning new ML techniques, we have learned to integrate them in a compact system (which is a huge challenge it itself)

What's next for Pointy

Since this project it a prototype, Pointy has huge potential in terms of it's potential growth and the number of people it can impact. Besides making the ML algorithms more efficient, Pointy will most likely benefit of better accessibility for it's target group (i.e. further miniaturization or recognition of many more object in context)

Built With

• python