-
Notifications
You must be signed in to change notification settings - Fork 1
Project vision
The goal of the project is to create a robot capable of recognizing people's faces and greeting them by name.
The robot will perform facial recognition using image data from a webcam and an external library, and compare the faces detected in the video frames to a set of known faces. Upon finding a match, it will greet the person by their name using a speech synthesizer and a connected speaker.
The robot should also be able to move around autonomously, using a laser scanner to detect and avoid obstacles. It should additonally be capable of detecting people standing nearby, and move towards the nearest person in order to greet them.
The robot will be used by Department of Computer Science of AGH University of Science and Technology to greet its students in the D17 building corridors.
- Recognizing when a person is in front of the robot
- Identyfing people based on their face
- Greeting people by saying their name
- Autonomous movement, avoiding obstacles
- Low latency, real-time recognition.
- Using Capo Robot base, running on PandaBoard
- Using facial recognition library provided by project supervisor. Library itself is written in C++, with wrappers for Java available.
- Set of reference photographs of students' faces, with at least 60 pixels between the eyes.
Performing facial detection/recognition for high-resolution images may take a long time. Additional preprocessing might be required in order to speed up the process. Alternatively, the recognition could be performed on a remote machine with more processing power, at the cost of portability (the system will no longer be self-contained)