Xu Chen
Research Subjects
Research Interests
Xu Chen has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Otmar Hilliges
ETH Zurich
2 shared publications - Thabo Beeler
Google, Zurich, Switzerland
2 shared publications - Abhimitra Meka
Google Inc., United States of America
1 shared publication - Andreas Geiger
University of Tübingen
1 shared publication - Ayce Idil Aytekin
Max Planck Institute for Informatics, Saarbruecken, Germany and Saarbruecken Research Center for Visual Computing, Interaction and Artificial Intelligence, Saarbruecken, Germany
1 shared publication - Christian Theobalt
Max Planck Institute for Informatics, Saarbrücken, Germany
1 shared publication
Latest Publications
- Open Access
- 30 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Grasp in Gaussians: Fast Monocular Reconstruction of Dynamic Hand–Object Interactions
- 30 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Open AccessGrasp in Gaussians: Fast Monocular Reconstruction of Dynamic Hand–Object Interactions
- Open Access
- 3 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-12
Cafca: High-quality Novel View Synthesis of Expressive Faces from Casual Few-shot Captures
- 3 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-12
Open AccessCafca: High-quality Novel View Synthesis of Expressive Faces from Casual Few-shot Captures
- 28 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 45 (10) , 11796-11809
Fast-SNARF: A Fast Deformer for Articulated Neural Fields
Neural fields have revolutionized the area of 3D reconstruction and novel view synthesis of rigid scenes. A key challenge in making such methods applicable to articulated objects, such as the human body, is to model the deformation of 3D locations between the rest pose (a canonical space) and the deformed space. We propose a new articulation module for neural fields, Fast-SNARF, which finds accurate correspondences between canonical space and posed space via iterative root finding. Fast-SNARF is a drop-in replacement in functionality to our previous work, SNARF, while significantly improving its computational efficiency. We contribute several algorithmic and implementation improvements over SNARF, yielding a speed-up of $150\times$ . These improvements include voxel-based correspondence search, pre-computing the linear blend skinning function, and an efficient software implementation with CUDA kernels. Fast-SNARF enables efficient and simultaneous optimization of shape and skinning weights given deformed observations without correspondences (e.g. 3D meshes). Because learning of deformation maps is a crucial component in many 3D human avatar methods and since Fast-SNARF provides a computationally efficient solution, we believe that this work represents a significant step towards the practical creation of 3D virtual humans.
- 28 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 45 (10) , 11796-11809
Fast-SNARF: A Fast Deformer for Articulated Neural Fields
Neural fields have revolutionized the area of 3D reconstruction and novel view synthesis of rigid scenes. A key challenge in making such methods applicable to articulated objects, such as the human body, is to model the deformation of 3D locations between the rest pose (a canonical space) and the deformed space. We propose a new articulation module for neural fields, Fast-SNARF, which finds accurate correspondences between canonical space and posed space via iterative root finding. Fast-SNARF is a drop-in replacement in functionality to our previous work, SNARF, while significantly improving its computational efficiency. We contribute several algorithmic and implementation improvements over SNARF, yielding a speed-up of $150\times$ . These improvements include voxel-based correspondence search, pre-computing the linear blend skinning function, and an efficient software implementation with CUDA kernels. Fast-SNARF enables efficient and simultaneous optimization of shape and skinning weights given deformed observations without correspondences (e.g. 3D meshes). Because learning of deformation maps is a crucial component in many 3D human avatar methods and since Fast-SNARF provides a computationally efficient solution, we believe that this work represents a significant step towards the practical creation of 3D virtual humans.
- 8 July 2020
- Published by ASME International
Active Interlayer Heating for Sustainable Selective Laser Sintering With Reclaimed Polyamide 12 Powders
Selective laser sintering (SLS) technology produces a substantial amount of un-sintered polyamide 12 powders after the manufacturing process. Failure to recycle and reuse these aged powders not only leads to economic losses but also is environmentally unfriendly. This is particularly problematic for powder particles close to the heat-affected zones that go through severe thermal degradations during the laser sintering processes. Limited procedures exist for systematically reusing such extremely aged powders. This work proposes a systematic method to maximize reusability of aged and extremely aged polyamide 12 powders. Building on a previously untapped interlayer heating, pre-processing, and a systematic mixing of powder materials, we show how reclaimed polyamide 12 powders can be consistently reprinted into functional samples, with mechanical properties even superior to current industrial norms. In particular, the proposed method can yield printed samples with 18.04% higher tensile strength and 55.29% larger elongation at break using as much as 30% of extremely aged powders compared to the benchmark sample.
- 8 July 2020
- Published by ASME International
Active Interlayer Heating for Sustainable Selective Laser Sintering With Reclaimed Polyamide 12 Powders
Selective laser sintering (SLS) technology produces a substantial amount of un-sintered polyamide 12 powders after the manufacturing process. Failure to recycle and reuse these aged powders not only leads to economic losses but also is environmentally unfriendly. This is particularly problematic for powder particles close to the heat-affected zones that go through severe thermal degradations during the laser sintering processes. Limited procedures exist for systematically reusing such extremely aged powders. This work proposes a systematic method to maximize reusability of aged and extremely aged polyamide 12 powders. Building on a previously untapped interlayer heating, pre-processing, and a systematic mixing of powder materials, we show how reclaimed polyamide 12 powders can be consistently reprinted into functional samples, with mechanical properties even superior to current industrial norms. In particular, the proposed method can yield printed samples with 18.04% higher tensile strength and 55.29% larger elongation at break using as much as 30% of extremely aged powders compared to the benchmark sample.