Dengsheng Chen
About Me
I earned my Master’s in Computer Science from the National University of Defense Technology in 2022. Following my graduation, I joined the Visual Intelligence Department at Meituan. My current work is centered on pioneering research in multi-modal understanding and generating images, videos, and 3D content.
Last updated: 2024.10.8
Education
- Master’s Degree in Computer Science, National University of Defense Technology, 2019-2022
- Bachelor’s Degree in Computer Science, Fuzhou University, 2015-2019
Experience
- Meituan, AI Researcher (2022.6~Present): AIGC: Image/Video/3D generation.
- ByteDance, AI Intern (2021.1~2021.12): Federated learning: Algorithm optimization and federated learning system.
- Ecovacs Nanjing AI Research Institute, AI Intern (2019.1~2019.8): SLAM: Dense indoor environment construction for sweeping robot.
- Tencent AI Lab, AI Intern (2018.6~2018.12): Depth fusion: 3D face reconstruction from iPhone depth camera.
Preprint & Publication
- Dengsheng Chen, Jie Hu, Xiaoming Wei, and Enhua Wu. Denoising with a Joint-Embedding Predictive Architecture. arXiv preprint arXiv:2410.03755, 2024. paper code homepage
- Dengsheng Chen, Jie Hu, Xiaoming Wei, and Enhua Wu. Fine-gained Zero-shot Video Sampling. arXiv preprint arXiv:2407.21475, 2024. paper code homepage
- Dengsheng Chen, Jie Hu, Xiaoming Wei, and Enhua Wu. Deformable 3D Shape Diffusion Model. arXiv preprint arXiv:2407.21428, 2024. paper
- Dengsheng Chen, Xiaoming Wei, and Xiaolin Wei. Animating general image with large visual motion model. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7131–7140, 2024. paper code homepage
- Dengsheng Chen, Jie Hu, Xiaoming Wei, and Enhua Wu. Real3d: The curious case of neural scene degeneration. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 1028–1036, 2024. paper
- Dengsheng Chen, Jie Hu, Vince Junkai Tan, Xiaoming Wei, and Enhua Wu. Elastic aggregation for federated optimization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12187–12197, 2023. paper code
- Dengsheng Chen, Vince Junkai Tan, Zhilin Lu, Enhua Wu, and Jie Hu. OpenFed: A comprehensive and versatile open-source federated learning framework. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5018–5026, 2023. paper code
- Dengsheng Chen, Jie Hu, Wenwen Qiang, Xiaoming Wei, and Enhua Wu. Rethinking skip connection model as a learnable Markov chain. In The Eleventh International Conference on Learning Representations. paper code
- “Improved Squeeze-and-Excitation Networks by Eliminating Disharmony from Collaboration with Normalization.” 2022.12
- Dengsheng Chen, Jun Li, and Kai Xu. Arelu: Attention-based rectified linear unit. arXiv preprint arXiv:2006.13858, 2020. paper code
- Dengsheng Chen, Haowen Deng, Jun Li, Duo Li, Yao Duan, and Kai Xu. Potential convolution: Embedding point clouds into potential fields. arXiv preprint arXiv:2104.01754, 2021. paper
- Dengsheng Chen, Jun Li, Zheng Wang, and Kai Xu. Learning canonical shape space for category-level 6d object pose and size estimation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11973–11982, 2020. paper code
- Wenxi Liu, Yibing Song, Dengsheng Chen, Shengfeng He, Yuanlong Yu, Tao Yan, Gehard P Hancke, and Rynson WH Lau. Deformable object tracking with gated fusion. IEEE Transactions on Image Processing, 28(8):3766–3777, 2019. paper
- Dengsheng Chen, Yuanlong Yu, and Xiang Gao. Semi-supervised deep learning framework for monocular visual odometry. 2019. paper code
- Dengsheng Chen, Yuanlong Yu, and Zhiyong Huang. Online memory learning for active object recognition. In 2019 IEEE International Conference on Robotics and Biomimetics (ROBIO), pages 2914–2919. IEEE, 2019. paper
- Dengsheng Chen, Wenxi Liu, You Huang, Tong Tong, and Yuanlong Yu. Enhancement mask for hippocampus detection and segmentation. In 2018 IEEE International Conference on Information and Automation (ICIA), pages 455–460. IEEE, 2018. paper