Eye.ai

• 

  face detection result 1

• 

  face detection result 2

Inspiration

We believe that navigation for the blind is critically important nowadays. As the navigation system becomes increasingly advanced and is equipped with the most cutting-edge technology, such as machine learning algorithms and a worldwide database. We believe that we should also improve the living space for the blind and make them walk out more safely and comfortably.

What it does

It is a web-based application that utilizes computer camera to snapshot images of the motion for the blind. Then, by using our face detection algorithms, it could identify objects and indicate the danger level by verbally uttering to the blind users.

How we built it

Our application is based on the Google Cloud Platform. By using COCO models in Tensorflow and testing them on Jupyter Notebook, we have enabled our face detection on the web interface. Then, we use javascript and python to communicate the results of the face detection and speak them for the blind users.

Accomplishments that we're proud of

We have successfully located a very efficient face detection algorithm and implement it on the images of all types. Therefore, we are able to analyze a variety of real-life situations and send feedback properly to the blind users.

What we learned

Google Cloud Platform is very easy to use and fully equipped with all the functionality that we have desired throughout this project. Also, Tensorflow and Jupyter Notebook provide us with efficient face detection, which is a key feature in our application.

Built With

• css3
• google-cloud
• html
• javascript
• python