SE 329, Project 1
We will create an application that helps a professor identify the students in a classroom. The application will use a camera and facial recognition to scan students individually and display the name and picture of each student to the professor as it recognizes them. If the application sees a student who hasn’t been identified, it gives the professor the opportunity to enter the student’s name so he or she will be identified in the future.
These first two pictures illustrate a user training a new face into the system.
The next pictures show the system recognizing that user and displaying his name on the screen.
Project Koh will be a cloud based software system that houses the server. A cloud application model allows for multiple front-end applications to interact with the server across multiple devices. This would provide professors many options to learn more about their students. However, the first versions will only be compatible with laptop computers with a camera.
Project Koh is what Iowa State University needs to create personal relationships between faculty and students. When this relationship is formed, students are more likely to attend class. ISU can expect to see an increase students, who graduate in four years or less; decrease in dropped classes; and an increase of students hired within six months after graduation. These reported statistics will align with university trends that students who attend class on a regular basis have higher Grade Point Averages (GPA) than those who don’t.
The university can look to extend Project Koh’s applications to giving teaching assistants access to the application for recitations, as well as Supplemental Instruction leaders to lead their sessions. We want to encourage strong relationships to expand Iowa State’s continuous learning environment.
Beyond the MVP, our stretch goal is to deliver an extended product that has all the functionality of the MVP, as well as some or all of the following:
- A fully featured UI.
- Increased integration, providing the ability to take a picture from the UI directly.
- Multi-face recognition in a single photo (possibly from a video stream)
- Maintainable class roster.
- Multi-platform support.
Multi-platform operability would be possible by structuring the server in a generic way. Any user only needs to send an image to the server and it responds with information. Structuring the server this way provides the freedom to develop the client-side software independently.
Continuous video stream inspection would allow the interface to be hands free. Rather than having to click to take a photo and send it to the server, the whole process is automatic. The camera is continuously running and each frame is checked to see if it contains a face. If so, identity information is displayed.
An example of a combination of these stretch goals could be implemented on the Microsoft Hololens. If an instructor is wearing a Hololens while instructing, the device could continuously be scanning the recognizable faces in the audience and adding a name icon above them in real time using augmented reality. If a professor wants to call on a student, they would have their name available to them.
An example of the class roster function would allow the professor to go about lecture as he or she normally would while the system automatically takes attendance. If there are any unknown students or absent students, it would ask the professor to confirm whether the student was present or not. An extension of this idea could be a professor logging in and selecting which class he is teaching and a corresponding roster being loaded. If any students are unrecognized or incorrectly identified, the professor could select them from the list to help retrain the server.
For more information regarding the title of this project, click here.
International Justice League of Super Acquaintances
- Brody Concannon
- Nathan Karasch
- Stefan Kraus
- Gregory Steenhagen