AEye - Artificial Intelligence for Eye Tracking Data: Deep Learning Methods for the Automated Analysis of Cognitive Processes.
The way we move our eyes is very informative about the (often unconscious) processes that unfold in our minds. Research from cognitive psychology and psycholinguistics has shown that eye movements not only reflect temporary psychological conditions of an individual such as mental fatigue, drowsiness, or the level of concentration, but also higher-level cognitive processes involved in the comprehension and production of language.
We are an interdisciplinary research group with backgrounds in Computer Science, Mathematics, Linguistics, and Cognitive Science.
Our goal is to develop machine learning methods for the analysis of eye tracking data in order to make inferences or predictions about the cognitive processes and psychological states of an individual. We are developing end-to-end-trainable deep learning architectures that can process the signal recorded from a video-based eye-tracking device. The application areas of our methods range from the eye tracking-based diagnosis of developmental language disorders (dyslexia and Specific Language Impairment SLI), adaptive e-learning (e.g., the automated assessment of reading comprehension, foreign language skills, mental fatigue or distraction), driver monitoring (e.g., detection of drowsiness or alcoholization) to biometric user identification.
- Shuwen Deng, Paul Prasse, David R. Reich, Sabine Dziemian, Maja Stegenwallner-Schütz, Daniel Krakowczyk, Silvia Makowski, Nicolas Langer, Tobias Scheffer and Lena A. Jäger. Detection of ADHD based on Eye Movements during Natural Viewing. ECML 2022.(Video). (Code). (Poster).
- Paul Prasse, David R. Reich, Silvia Makowski, Lena A. Jäger and Tobias Scheffer. Fairness in Oculomotoric Biometric Identification. ETRA 2022. (Video). (Code)
- David R. Reich, Paul Prasse, Chiara Tschirner, Patrick Haller, Frank Goldhammer and Lena A. Jäger. Inferring Native and Non-Native Human Reading Comprehension and Subjective Text Difficulty from Scanpaths in Reading. ETRA 2022. (Video). (Code)
- Nora Hollenstein, Federico Pirovano, Ce Zhang, Lena Jäger and Lisa Beinborn. Multilingual Language Models Predict Human Reading Behavior. NAACL 2021.
- Paul Prasse, Lena A. Jäger, Silvia Makowski, Moritz Feuerpfeil, Tobias Scheffer. On the Relationship between Eye Tracking Resolution and Performance of Oculomotoric Biometric Identification. KES 2020. (Online talk)
Check out our code/software on github: https://github.com/aeye-lab