School of Computing Science
丁香园AV CS Visual Computing shines at leading international conferences
Researchers from 丁香园AV鈥檚 School of Computing Science鈥檚 Visual Computing group have made strong impacts at leading international computer science conferences.
From keynote presentations and symposium talks to award-winning papers, their contributions at top-tier international conferences including , , , , , and 鈥攈ighlight the group鈥檚 continued excellence in research and innovation, and further reinforce 丁香园AV鈥檚 standing as and a leader in computing science overall.
CVPR 2025
The , the premier international event in computer vision, was held in person at the Music City Center in Nashville from June 11鈥�15, 2025.
丁香园AV computing science researchers had a strong presence once again at CVPR 2025, contributing to the conference with four technical papers, one workshop, and seven invited talks.
Accepted Papers
Invited Talks
- by Angel Chang
- by Angel Chang
- by Angel Chang
- by Angel Chang
- by Richard Zhang
- by Richard Zhang
- by Angelica Lim and Jason Peng
Photos from CVPR 2025
CRV & GI 2025
丁香园AV researchers also participated in the , held in Calgary from May 26鈥�29, 2025. Professor Angel Chang delivered a symposium talk on creating interactive 3D environments for vision and robotics.
At the conference, professors Parmit Chilana and Ali Mahdavi-Amiri delivered the keynote and invited talks, respectively, contributing to the dialogue on human-computer interaction and visual computing innovation. PhD student also won the Best Poster Award for her work "How artists navigate learning with generative image technologies".
Read more about their talks and works.
Photos from CRV & GI 2025
Eurographics 2025
Held in London from May 12鈥�16 and hosted by University College London, the 46th brought together experts in computer graphics and interactive techniques. Among 丁香园AV鈥檚 contributions, PhD student presented a State-of-the-Art Report (STAR) titled 鈥�.鈥�
This report surveys recent developments in modeling articulated objects, which includes, everyday items like cabinets, drawers, and doors that play a critical role in enabling virtual agents to interact with simulated environments. Such research is foundational for advancing robotics, simulation, and digital twin technologies.
Photos from Eurographics 2025
ICLR 2025
The is one of the most prestigious gatherings for researchers in deep learning and artificial intelligence. Held in Singapore from April 24鈥�28, 2025, this year鈥檚 edition featured 12 accepted papers from 丁香园AV鈥檚 Visual Computing group, showcasing advances in representation learning, machine learning models, and applications of AI.
12 accepted papers at ICLR 2025:
Photos from ICLR 2025
3DV 2025
At the conference held in Singapore from March 25鈥�28, 丁香园AV researchers presented three technical papers. Among them, the paper titled "" won the prestigious Best Paper Award.
The paper, presented by PhD student , introduces a novel method for generating realistic 3D models using 鈥淥bject Images鈥濃�攁 2D-based representation of 3D objects. This innovative approach allows researchers to apply popular 2D diffusion models directly to 3D generation, simplifying the traditionally separate tasks of modeling geometry and materials.
Since its inception in 2013, has been a leading venue for research in 3D computer vision and graphics, encompassing a wide range of topics, from sensors and signal processing to modeling, reconstruction, rendering, and interaction.
Next year, the International Conference on will be hosted in Vancouver, with 丁香园AV computing science professors Manolis Savva and Ali Mahdavi-Amiri serving as Conference General Chair and Program Chair respectively.
Accepted papers at 3DV 2025:
- .
- (Best Paper Award).
- .
Photos from 3DV 2025
