Dr. Marco Pesavento

I am a Postdoctoral Research Fellow at the Centre for Vision, Speech and Signal Processing, University of Surrey, Uk. My current research is on improving the realism of 3D humans reconstructed from images. I am working on the AI4ME in collaboration with the BBC R&D under the supervision of Prof. Adrian Hilton, Dr. Armin Mustafa and Dr. Marco Volino.

I obtained my Ph.D. degree in Computer Vision and Graphic in January 2024 at the Centre for Vision, Speech and Signal Processing, University of Surrey, Uk. The Ph.D. thesis was on shape and appearance super-resolution of 3D humans under the supervision of Prof. Adrian Hilton and Dr. Marco Volino and explore novel deep learning approaches to reconstruct highly realistic 3D humans from a minimal number of images. During my Ph.D. I worked as a Research Scientist intern at Meta Reality Labs in San Francisco, US from June 2022 until November 2022 on 3D human reconstruction from RGB-D images.

Prior to this experience, I obtained my M.Sc. and B.Sc. degrees from the Information Engineering Department, DEI, at the University of Padova, Italy. In 2017, I was awarded with Erasmus+ scholarship that allowed me to complete the M.Sc. at the Department of Informatic, IFI of University of Oslo during which I collaborated with Simula Research Laboratory for my master dissertation.

I have particular interest in applying computer vision and machine learning to reconstruct, edit and render static and dynamic humans for application in film, broadcast and gaming. I am investigating novel representations that exploit deep learning to learn representations of people for highly realistic rendering. My research objective is to develop novel methods that reconstruct highly realistic digital humans from a minimum number of cameras to improve the content production for both traditional media film/games/TV and new forms of immersive augmented and virtual reality experience.

	COSMU: Complete 3D human shape from monocular unconstrained images M. Pesavento, M. Volino and A. Hilton European Conference on Computer Vision (ECCV 2024) Paper
	ANIM: Accurate Neural Implicit Model for Human Reconstruction from a single RGB-D image M. Pesavento, Y. Xu, N. Sarafianos, R. Maier, Z. Wang, C. Yao, M. Volino, E. Boyer, A. Hilton and T. Tung Proceedings in the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2024) Paper Project Page
	Super-resolution 3D Human Shape from a Single Low-Resolution Image M. Pesavento, M. Volino and A. Hilton European Conference on Computer Vision (ECCV 2022) Paper Project Page
	Image Super-Resolution via Hierarchical Attention-Based Multi-References Sampling M. Pesavento, M. Volino and A. Hilton The 18th ACM SIGGRAPH European Conference on Visual Media Production (CVMP 2021, Short) Paper (Best Short Paper Award)
	Attention-based Multi-Reference Learning for Image Super-Resolution M. Pesavento, M. Volino and A. Hilton Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV 2021) Paper
	Super-Resolution Appearance Transfer for 4D Human Performances M. Pesavento, M. Volino and A. Hilton Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Dynavis workshop (CVPRW 2021) Paper
	Effectiveness of VR immersive applications for public speaking enhancemen A. Notaro, F. Capraro, M. Pesavento, S. Milani and M.G. Busà Electronic Imaging 2021 (EI 2021) Paper
	Super-Resolution Appearance Transfer for 4D Human Performance Capture M. Pesavento, M. Volino and A. Hilton The 17th ACM SIGGRAPH European Conference on Visual Media Production (CVMP 2020, Short) Paper
	XR-Cockpit: a comparison of VR and AR solutions on an interactive training station A Caputo, S Jacota, S Krayevskyy, M Pesavento, F Pellacini and A Giachetti 25th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA 2020) Paper