Nedko Savov
Computer Vision / Machine Learning Specialist

I am a machine learning professional, passionate about computer vision. I work at 3DUniversum, where, in addition to other projects, I currently handle and develop the full process of generating high quality deepfakes - from the studio recording and the deep learnig pipeline to the video editting process. In the past 3 years, I have taken charge on projects on novel viewpoint generation, semantic segmentation of indoor scenes, face relighting, face analysis models (age and emotion recognition). I provided supervision for UvA Master AI students for their thesis projects on computational photography, optical flow, deepfake detection, autonomous driving.
Previously, I finished the Artificial Intelligence Master’s at the University of Amsterdam with honors. My thesis studies the benefits of using features rich on 3D facial information for age estimation. Before that, I was a student at Sofia University, where I completed my Computer Science Bachelor’s degree. There I lead practical sessions in teaching students as part of the course Objective Oriented Programming. In the past, I’ve participated in national and international programming contests as part of competetive programming courses.
publications
-
Spatio-temporal Features for Generalized Detection of Deepfake VideosarXiv preprint arXiv:2010.11844 2020
-
Pose and expression robust age estimation via 3D face reconstruction from a single imageIn Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops 2019