LookingGlass
Juan Sebastian Angarita | Vinay Vemuri | Varun Kohli | Edward Sears
LookingGlass is a window to your day allowing you to see your news, email, football stats, and other passions, as you brush your teeth, uninterrupted. By using facial recognition, the smart mirror will display your configured widgets and provide interaction by detecting basic hand gestures.
Why make something so cool?
We are in a race to the future. Smartphones are more powerful than computers that filled up entire rooms, robots visit completely alien worlds, nanoscale structures behave like machines, yet our homes remain in the past, static. LookingGlass is one of the first steps to making the spaces where we live truly futuristic.
Requirements
Functional
- Provides functionality of a basic mirror (reflect)
- Detect presence of individual in front of mirror and recognize registered users
- Present user-specific and user-agnostic widgets
- Allow basic interaction with widgets
- Allow new user registration and widget configuration
Non-Functional
- Expected to provide some functionality even if failures occur in parts of the product. If product fails because of environmental conditions (insufficient lighting), provide user feedback.
- Expected to recognize the presence of multiple individuals in viewport and hide private information (email) depending on user configuration.
- Expected to detect and recognize faces within 4 seconds.
- Must be designed to support a subset of standard commercial mirror sizes.
- Expected to support facial detection and recognition for a wide range of face shapes and sizes.
The Competition
Product |
Release Date |
Target Market |
Price Point |
Application |
Disney Smart Home |
Never |
Commercial |
None |
Commercial Display |
Specular.tv Smart Mirror |
Per Contract |
Commercial |
Per Contract |
Commercial Display |
Seraku Smart Mirror |
TBD |
Commercial |
Per Contract |
Commercial Display |
NYT Magic Mirror |
Never |
Commercial |
None |
Novelty |
Seura TV Mirror |
Now |
Consumer |
$2699 |
TV Mirror |
Cybertecture Mirror |
Now |
Commercial |
$7700 |
Personalized Display |
LookingGlass |
3 Years |
Consumer |
$450 |
User-centric extensible display |
Risks & Mitigations
Risk |
Mitigation |
Do not know facial detection process intensity |
Run benchmarks on different processors |
Do not know best facial recognition algorithms |
Consult CMU professors |
Not hiding personal information quickly enough |
Optimize facial detection algorithm |
Test Cases
Name |
Why |
How |
Faults |
No user in front of mirror |
There might be no user in front of the mirror. |
No user in front of mirror. |
No artificial faults |
Number of simultaneous users |
There might be multiple users in front of the mirror at one time. |
Multiple users in front of mirror. |
No artificial faults |
Not registered user |
There might be a non registered user in front of the mirror. |
Unregistered user in front of mirror. |
No artificial faults |
Brief exposure to user |
A user might only be in front of the mirror for a brief period of time. |
User in front of mirror for brief period of time. |
No artificial faults |
Multiple sensor triggers at once |
A user might trigger multiple sensors at a time. |
Trigger multiple sensors at a time. |
No artificial faults |
Trigger a sensor when no user in front of mirror |
A user might trigger a sensor accidentally while not in front of mirror. |
Trigger a sensor with no user in front of mirror. |
No artificial faults |