Addressing visual bias
- Ioannis Sarridis, Deep Learning Researcher (CERTH)
- Dr Elmokhtar Mohamed Moussa, Postdoctoral Researcher on Deep learning & Computer vision (IDnow)
- Dr Miquel Pérez Torres, Educational Developer (University of Groningen)
- Dr Francisco Hernández, Postdoctoral Researcher and Lecturer (University of Groningen)
Webinar objectives
🎯
- Explore Three Methods for Mitigating Visual Biases
- Enhance Identity Verification with Synthetic Data Augmentation
-
- Discuss challenges in selfie-to-ID matching due to the lack of documents specific visual features in the public datasets and from skin tone biases, and present a framework for generating synthetic ID documents.