Newswise — New York, June 9, 2026 – The IEEE Computer Society (CS) and the Computer Vision Foundation (CVF) recognized outstanding achievements in computer vision with the award-winning papers of the 2026 Conference on Computer Vision and Pattern Recognition (CVPR).
Best paper award
After a rigorous review process in which 4,089 papers were accepted from 16,092 submissions, the CVPR 2026 Award Selection Committee selected the following two papers for top honors at this year’s conference:
CVPR 2026 Best Paper
- Efficiently reconstruct dynamic scenes one D4RT at a time, author: Zhang Chuhan. Guillaume Le Moyne; Skanda Koppula. Ignacio Rocco; Lilian Momeni. Xie Runyu. Sun Xiuyang. Rahul Sukhtankar. Joel K. Barral. Lia Hadsell. Zubin Ghahramani;Andrew Zisserman;Junlin Zhang;Medi SM Sajadi – A team from Google DeepMind, University College London, and the University of Oxford has developed D4RT, a network that can reconstruct the geometry and motion of dynamic 4D scenes from video. This model uses an integrated transformer-based architecture to estimate depth, spatiotemporal correspondence, and complete camera parameters, allowing the 3D location of any point in space and time to be independently and efficiently investigated. D4RT provides a lightweight and highly scalable method for highly efficient training and inference by simplifying previously computationally intensive processes.
CVPR 2026 Best Student Paper
- Native compact structured latent for 3D generation, Author: Xianfeng; Xiaoxue Chen. Xu Sicheng. Wang Ruicheng; Zelong Lv; Deng Yu Zhu Hongyuan. Yuedong. Hao Zhao. Nicholas Jin Yuan. Yang Jiaolong – A team from Tsinghua University, Microsoft Research, University of Science and Technology of China, and Microsoft AI has developed a new approach to 3D generative modeling that significantly improves the quality and realism of AI-generated 3D assets. The research focuses on O-Voxel, a new representation that can accurately capture complex shapes and surface attributes. O-Voxel represents a significant advance in 3D generative modeling, with the geometry and quality of generated assets far exceeding existing models.
“This year’s winning papers demonstrate the innovation and technical excellence that continues to move the field forward,” said Alexander G. Schwing, associate professor of electrical and computer engineering at the University of Illinois at Urbana-Champaign, Illinois, and co-chair of the CVPR 2026 program. “From advances in dynamic scene reconstruction to breakthroughs in 3D generative modeling, these works address fundamental challenges in computer vision while opening up new possibilities for applications in AI, robotics, and more. We congratulate the authors on this well-deserved recognition and thank the awards committee for their thoughtful evaluation process.”
In addition to these top papers, the awards committee identified two Best Paper Honorable Mentions and one Student Paper Honorable Mention.
CVPR 2026 Best Paper Honorable Mention
- NitroGen: An Open Foundation Model for Generalist Game Agents, Authors: Loïc Magne, Anas Awadalla, Guanzhi Wang, yingzhen Xu, Joshua Belofsky, Fengyuan Hu, Joohwan Kim, Ludwig Schmidt, Georgia Gkioxari, Jan Kautz, Yisong Yue, Yejin Choi, Yuke Zhu, Linxi Fan – A team including authors from NVIDIA, Stanford University, California Institute of Technology, University of Chicago, and University of Texas at Austin introduced NitroGen, a vision-action foundation model for generalist game agents. It has been trained with 40,000 hours of gameplay videos across over 1,000 games, giving you powerful abilities across a variety of areas.
- SAM 3D: 3Dfy everything in your images, Authors: Xingyu Chen, FU-JEN CHU, Pierre Gleize, Kevin J Liang, Alexander Sax, Hao Tang, Weiyao Wang, Michelle Guo, Thibaut Hardin, Xiang Li, Aohan Lin, Jia-Wei Liu, Ziqi Ma, Anushka Sagar, Bowen Song, Xiaodong Wang, Jianing Yang, Bowen Zhang, Piotr Dollar, Georgia Gukiosari, Matt Faizuli, Jitendra Malik – The team at Meta Superintelligence Labs announced SAM 3D, a generative model for predicting geometry, texture, and layout from a single image and reconstructing visually grounded 3D objects. This work is a significant improvement over recent works, showing at least a 5:1 win ratio on tests of human preference regarding real-world objects and scenes.
CVPR 2026 Best Student Paper Honorable Mention
- ChordEdit: One-step low-energy transfer for image editing, Authors: Liangsi Lu, Xuhang Chen, Minzhe Guo, Shichu Li, Jingchao Wang, Yang Shi – A team from Guangdong University of Technology, Huizhou University, Shenzhen University, and Peking University introduced ChordEdit, a model-agnostic, no-training, no-inversion method that facilitates high-fidelity one-step image editing. It enables fast, lightweight, precise editing, and ultimately true real-time editing of even the most difficult models.
All award-winning papers demonstrate impressive results that help advance computer vision, artificial intelligence (AI), and more. These were announced at the conference welcome and awards ceremony on June 5th and presented at the IEEE Computer Society Technical Community on Pattern Analysis and Machine Intelligence (TCPAMI) conference on June 6th.
“These papers represent some of the most impactful and forward-looking research presented at CVPR 2026,” said Chen Cheng Loi, Tan Lip Bu AI Professor in the School of Computing and Data Science at Nanyang Technological University in Singapore and CVPR 2026 Program Co-Chair. “The selected papers introduce concepts and practices that will undoubtedly help shape the next generation of computer vision systems and applications.”
About CVPR 2026
The Conference on Computer Vision and Pattern Recognition (CVPR) is the premier computer vision event for emerging research supporting artificial intelligence (AI), computer vision, multimodal AI, wearable AI, new AI architectures, spatial computing, agent AI, scientific AI, embodied AI and robotics, and more. Sponsored by the IEEE Computer Society (CS) and the Computer Vision Foundation (CVF), CVPR delivers important advances in computer vision and pattern recognition, expanding the frontiers of computer vision and defining the next generation of intelligent, interactive technologies. With a top-notch technical program including tutorials and workshops, cutting-edge immersive exhibitions, and rich networking opportunities, CVPR attracts more than 10,000 scientists and engineers from around the world each year, creating unique opportunities for networking, recruiting, inspiration, and motivation.
CVPR 2026 was held June 3-7 at the Colorado Convention Center in Denver, Colorado, USA. CVPR 2027 was held June 19-26 at the Seattle Convention Center in Seattle, Washington, USA. For more information about CVPR, please visit cvpr.thecvf.com.
About the Computer Vision Foundation
The Computer Vision Foundation (CVF) is a nonprofit organization dedicated to promoting and supporting research in all aspects of computer vision. Together with the IEEE Computer Society, we co-sponsor two of the largest computer vision conferences: CVPR and the International Conference on Computer Vision (ICCV). For more information, please visit thecvf.com.
About the IEEE Computer Society
The IEEE Computer Society engages computer engineers, scientists, academia, and industry experts from all areas of computing to set standards of excellence and advance global progress for the benefit of humanity.
The IEEE Computer Society will celebrate its 80th anniversary in 2026 and continues to build on its rich heritage. Through conferences, publications, and programs that bring together computer science and engineering leaders at all stages of their careers, the IEEE Computer Society supports, shapes, and guides the future of more than 375,000 community members and the larger computing community, enabling new opportunities to better serve the world. For more information, visit computer.org.
