Embedded Files
Eric Brachmann is a senior staff scientist at Niantic. He is well-known for his pioneering work on efficient and scalable scene coordinate regressors (SCRs) for localization and mapping. SCRs are considered privacy-preserving by many in the community, as it seems hard to extract the geometry from their neural network-based scene representations. Eric’s talk will provide an interesting perspective on SCRs in the context of privacy-preserving localization and mapping.
Webpage: https://ebrach.github.io/
Gabriela Csurka is a Principal Scientist at NAVER LABS Europe, France. Her main research interests are image and 3D understanding, visual localization, domain adaptation, and visual language models. She has a strong expertise in learning scene representations that facilitate accurate camera pose estimation for (privacy-preserving) visual localization. Gabriela will present her work on using (semantic) image segmentation for privacy-preserving visual localization.
Webpage: https://europe.naverlabs.com/people_user_naverlabs/Gabriela-Csurka/
Je Hyeong Hong is an assistant professor in the Department of Electronic Engineering at Hanyang University in Seoul. He and his group are actively working on geometric obfuscation schemes (both in image and 3D scene space) for privacy-preserving localization. Thus, his talk will provide an interesting perspective on such approaches.
Webpage: https://sail.hanyang.ac.kr/members
Marc Pollefeys is a Professor of Computer Science at ETH Zurich and the Director of the Microsoft Mixed Reality and AI Lab in Zurich. He is well-known for his groundbreaking work on 3D computer vision. He is one of the pioneers who introduced the privacy-preserving localization and mapping problems to the community. He and his group invented the concepts of geometry and descriptor obfuscation. His talk will provide an overview over state-of-the-art obfuscation-based approaches as well as a historical perspective.
Google Sites
Report abuse