Though webcams (and front-facing cameras in general) have become better over the past years due to increasing demand for personal promotion on social media, they are mostly unsuitable for eye tracking. The reason being that images do not contain well defined iris edges and eye corners from which traditional eye tracking is often performed.
This project aims to develop a pipeline which performs reliable and reasonably accurate eye tracking in real-time for consumer front-facing cameras, potentially extending to mobile devices.
You will work on existing code, incorporating my gaze estimation algorithm into your pipeline.
- Image processing
- Computer vision
- Much enthusiasm
Experience desired in:
- Deep Learning
- Training of CNNs
- Mobile app development
Implement a pipeline for eye tracking (face detection -> face segmentation -> facial-landmarks estimation -> eye segmentation -> gaze estimation) to potentially run real-time on consumer devices with front-facing cameras.
Seonwook "Wookie" Park (email@example.com)
IDEA League Student Grant (IDL)
CLS Student Project (MPG ETH CLS)
Information, Computing and Communication Sciences