Anonymous Professional Services
|
Next Challenge: What is your idea for a sustainable and cost-effective alternative to mineral oil? |
Our company has developed a facial recognition algorithm that is used to identify faces from video images automatically. Here's how it works...
- Video stills are analyzed and faces are marked with bounding boxes.
- The facial nodes on each of the faces are identified, creating a unique “finger print” for each of the detected face(s).
- The detected face(s) are compared to digital photo(s) stored in a dedicated database.
- The user should be presented with the result based on two potential use cases:
A) Is the face captured the owner of the card used to access the premises? The tools will compare the face detected to a single photo and output the % probability that it is the same person; this helps verify that the person requesting access is actually the person he/she claims to be.
B) Does the face match any of the individuals in the reference database? The tools will compare the face detected to multiple photos in a database and output the top matches with % probability for each match (i.e. how likely that it is the person in the video).
Your task is to design an interface for a technical system admin who is testing out the flow/accuracy of this facial recognition algorithm. Please design specifically for desktop but we would love to see a mobile dashboard concept screen as a bonus. Your desktop screens (jpg or png) should enable the user to:
-
See results
- Review whether there is a face match
- View cropped, zoomed photo with facial-nodes next to matched database photo
-
Improve the system and train the algorithm
- Adjust or add bounding boxes, adjust or add facial nodes
- Tell the algorithm when it is wrong (e.g. for use case B above, user could be presented with top 5 matches and % probability for each to select which is best).
-
Manage the application
- Add, remove, and review entries from the reference database
- Search the database and see recognition statistics about the face on the video (specific score with face detected on the screen)
- Review video recordings related to specific face recognition events (rewind, playback, etc.)
- Adjust settings such as: thresholds for flagging matches in live video etc. (for use case B)
The color scheme for our business is red, white, and blue.
***Feedback will be given to designs which are submitted by end of day Friday. We highly encourage submitting early so that you can make adjustments that increase your chances of winning.
Final screens *MUST* to be added to an InVision prototype or similar program. Submit a mobile concept summary screen for bonus points.
-
Submissions will be graded on the following criteria:
- Wow Factor
- Sleek/Simple
- Usability
will receive $1,000
will receive $600 each
| $1,000.00 |
|
Jennifer Whyte Savannah College of Art and Design | |
| $600.00 |
|
Fujie Zong Savannah College of Art and Design | |
| $600.00 |
|
Tin Le University of Oregon | |
| $600.00 |
|
Suleiman Shakir Rutgers, The State University of New Jersey |