Atheia

The Next-Gen Visual Aid

Innovative Features

Facial Recognition

Recognizes faces of people you interact with everyday.

Text Reader

Scan text from anywhere and have it read to you.

Search Mode

Actively search for objects in your vicinity and know how far they are.

Real-Time

Results are produced within seconds through our onboard processing.

Real-Time Stereo Processing

Our app can process a stereo live stream in real-time allowing us to actively locate objects in 3D space. We do this by creating a depth map, which can retrieve the distances to objects in the picture, which is key to our search mode.

On Board Processing

Processing is done locally on the device instead of through cloud services to get faster and more reliable coverage anywhere you go.

Long Battery Life

Our low power and highly efficient ARM processor can run for extended periods of time under high load.

GPS Tracking

Using an on-board GPS, Atheia tracks which locations the user frequents. Using this information, we say where specific objects were last seen.

Haptic Feedback

Depth information captured by the phone is conveyed through varying intensity of haptic feedback, giving the user an additional level of spatial understanding.

Easy Setup

The app requires little to no initial setup, however adding faces for facial recognition will require some manual input through the app.

Voice Assistant

All commands can be issued via voice commands, simplifying the operation of the device and allowing for a hands-free operation to maximize the ease of use.

State of the Art Computer Vision

We use a combination of state of the art computer vision models, optimized to run on our device. Through our custom algorithms, we are able to increase the accuracy of the classifications by deploying machine learning techniques.

Demo

Scan

Returns a list of detected objects and their frequency, sorted by closest to farthest:
12 people, 1 handbag, 1 bottle, 5 chairs, 1 table, and 3 laptops

Text

Retrieves all the text and states which objects the text is on:
The person says nc state and adidas

Facial Recognition

Recognizes specific people that the user can set up. Returns people and their relative position to you:
Michael is to the left, Dan is in the center, Martha is to the right

Search

Vibrates device in the direction of an object or describes the location: Search for bottle
Classifications overlaid onto depth map

Last Seen

Gives the time, location, and surrounding items of the specified object: Find 'calculator'
Calculator last seen next to 1 chair, 1 table, 4 people at 12:34 PM on June 24 at Home.

Smart Question

Answers specific questions about the environment: What is the color of the person's jacket?
The color is red


Prototypes

V7

V6

V5

V4

V3

V2

V1

Our Timeline

Prototyping

Sep - March

App Development

December - April

Small Scale Interview

May

Small Scale Testing

June

Beta Development

July - December

Large Scale Testing

January - Jun

Final Development

July - December

Alpha Release

March



Partners


Meet The Team

We are a group of high school students hoping to make lives better.

Eugene Choi

Co-Founder

Pranav Ravella

Co-Founder

Irfan Nafi

Co-Founder

Raffu Khondaker

Co-Founder