Register now After registration you will be able to apply for this opportunity online.
3D Human Modeling and Performance Capture
Digital capture of human bodies is a rapidly growing research area in computer vision and computer graphics that puts scenarios such as life-like mixed-reality (MR) virtual-social interactions into reach. Therefore, we offer projects for modeling and capturing humans at the intersection of computer vision, computer graphics, and machine learning.
Keywords: computer vision, deep learning, 3D reconstruction, 3D human
This project can be divided into multiple sub-topics, each of which can serve as a semester or thesis project. Our general goal for these projects is to push the next generation of full-body 3D human avatar modeling and intricate human performance capturing from videos. This will include but is not limited to neural implicit human representation, Gaussian splitting for avatars, and data-driven manners for learning large human models that allow high-quality and efficient human reconstructions. Here are two related works from our group ( https://moygcc.github.io/vid2avatar/ , https://skype-line.github.io/projects/X-Avatar/ ). Both have polished code bases for further build-on. You're welcome to reach out for more project details and technical overviews.
This project can be divided into multiple sub-topics, each of which can serve as a semester or thesis project. Our general goal for these projects is to push the next generation of full-body 3D human avatar modeling and intricate human performance capturing from videos. This will include but is not limited to neural implicit human representation, Gaussian splitting for avatars, and data-driven manners for learning large human models that allow high-quality and efficient human reconstructions. Here are two related works from our group ( https://moygcc.github.io/vid2avatar/ , https://skype-line.github.io/projects/X-Avatar/ ). Both have polished code bases for further build-on. You're welcome to reach out for more project details and technical overviews.
A high-fidelity and fully expressive human model that can be captured from images/videos via our proposed reconstruction approach in an efficient way. Eventually, the project work is targeted to be published in top-tier computer vision conferences or journals.
A high-fidelity and fully expressive human model that can be captured from images/videos via our proposed reconstruction approach in an efficient way. Eventually, the project work is targeted to be published in top-tier computer vision conferences or journals.