Shangchen Zhou
Shangchen Zhou is an individual computer-vision and generative-AI researcher and solo creator of an open research portfolio. His work focuses on multimodal generation and editing, agentic video generation, interactive world models, spatial intelligence, visual restoration, video inpainting, enhancement, and related datasets.
At a Glance
- Researchers and universities
- Open-source AI and computer-vision developers
- Media, entertainment, advertising, and visual-effects practitioners
- Archives and organizations restoring historical visual content
- +1 more
AI Tools by Shangchen Zhou
(1)ProPainter
Video Inpainting Model for Object Removal
Discussions
No discussions yet
Be the first to start a discussion about Shangchen Zhou
Latest News
Released Mundo, a customized world model that turns an environment into a programmable world.
Serving as an Area Chair for NeurIPS 2026, CVPR 2027, ICLR 2027, AAAI 2027, ICML 2027, and WACV 2027.
Received the NTU CCDS Outstanding PhD Thesis Award for 2024.
Appointed Research Assistant Professor at Nanyang Technological University.
Products & Services
Blind face-restoration algorithm for enhancing old photos, videos, and AI-generated faces. Its Transformer-based restoration uses a learned codebook and controllable fidelity-quality tradeoff; code, Hugging Face, Replicate, OpenXLab, and Colab access are linked from Zhou's site.
Video-inpainting and object-removal framework supporting object removal, video completion, and video outpainting. It combines recurrent flow completion, dual-domain image/feature propagation, and a mask-guided sparse video Transformer.
Text-guided diffusion model for temporal-consistent real-world video super-resolution and enhancement of real-world and AIGC videos.
Diffusion model for enhancing real-world and AI-generated images.
Market Position
This is an academic/open-source research portfolio rather than a commercial company. Its projects compete technically with prior and alternative computer-vision methods in video inpainting, face restoration, super-resolution, matting, and generative video, while differentiating through public code, datasets, project pages, hosted demos, and reported adoption/licensing by more than 20 companies and institutions.
Leadership
Executive Team
Shangchen Zhou
Postdoctoral Researcher / Research Associate in Multimodal Visual Editing, Imperial College London
Currently in Imperial College London's Department of Computing. Previously Research Assistant Professor at MMLab@NTU, Nanyang Technological University (2025–2026), and PhD researcher at MMLab@NTU/NTU advised by Professor Chen Change Loy; received the NTU CCDS Outstanding PhD Thesis Award.
Founding Story
This is not a company: Shangchen Zhou's public portfolio grew out of his academic research in computer vision, including work at MMLab@NTU and Nanyang Technological University. He publishes research papers, datasets, project websites, and open-source implementations so that visual restoration, editing, generation, and enhancement methods can be used and extended by researchers and developers.
Target Markets
- Researchers and universities
- Open-source AI and computer-vision developers
- Media, entertainment, advertising, and visual-effects practitioners
- Archives and organizations restoring historical visual content
- Photographers and creators working with degraded or low-light imagery
- Academic computer-vision research and benchmarking
- Developer experimentation and open-source model integration
- Video object removal, watermark removal, completion, and outpainting
- Restoration of old photos, films, family media, and historical documents
- Entertainment, advertising, archival preservation, and brand-content enhancement
- Low-light photography and real-world image/video enhancement