CVPR 2025 Showcases Future of AI and Computer Vision

The IEEE Computer Society (CS) and the Computer Vision Foundation (CVF) have reported that the 2025 Conference on Computer Vision and Pattern Recognition (CVPR) closed with “record-breaking technical success”.

With the highest number of paper submissions ever – 12,008, a 13 percent increase over 2024 – and 9,375 registrants from 75 different countries and regions, CVPR provided an environment ripe for discovery, collaboration, and advancement.

“CVPR continues to set the bar for research in computer vision,” said Bryan Morse, CVPR 2025 General Co-Chair and Professor of Computer Science at Brigham Young University in Provo, Utah, US. “We can expect the research from this year’s event to be cited into the future.”

Technical Program Highlights

This year’s conference, which took place 11-15 June in Nashville, Tennessee, US, convened the global computer vision community for the latest advances in the field.

CVPR 2025’s largest topic areas included image and video synthesis; 3D from multi-view and sensors; multi-modal learning; human face, body, pose, gesture, and movement; low-level vision; and vision language and reasoning.

With CVPR proceedings having earned the number two spot in Google’s 2024 Scholar Metrics, this year’s accepted papers likely will continue that trajectory, with the conference remaining a highly competitive and citable technical event.

In fact, only 22.1 percent of all submitted papers were accepted for presentations, and of those, a mere 3.3 percent were chosen as oral presentations.

In addition, the 118 workshops and 25 tutorials offered deeper reviews of prominent use cases for computer vision from autonomous vehicles to medicine, robotics, and more.

“As the field advances, we are seeing more and more applications of computer vision to other disciplines too,” said Phillip Isola, CVPR 2025 program co-chair and an associate professor at the Massachusetts Institute of Technology (MIT) in Boston, Massachusetts, US.

“The new methods shown at the conference will likely influence many research communities beyond just computer vision.”

Celebrating Top Research: Paper Award Winners

For example, advancements like those demonstrated in the conference best papers will shape what is to come.

Of the 2,872 accepted papers presented at CVPR 2025, two were singled out as top papers:

• VGGT: Visual Geometry Grounded Transformer, which introduces a feed-forward neural network that can directly estimate all key 3D scene properties for hundreds of input views, outperforming standard approaches, as Best Paper; and
• Neural Inverse Rendering from Propagating Light, which demonstrates a method to model and invert multi-viewpoint, time-resolved measurements of propagating light from a LiDAR system to recover scene geometry and to render videos of propagating light, as Best Student Paper.

“All of the papers accepted to CVPR help to advance the work of the community,” said Fuxin Li, CVPR 2025 program co-chair, and an associate professor at Oregon State University in Corvallis, Oregon, US.

“It will be exciting to watch how research continues to unfold over the next year.”

Insights from Industry Leaders: Keynote Highlights

Providing insights into significant topics for the industry, CVPR 2025 keynotes addressed how artificial intelligence (AI) can support the low-altitude economy; the next phase of AI; and the evolution of robotics with the support of more advanced AI. Specifically, the following presenters spoke to these themes in their talks:

• Harry Shum, former executive vice president of Microsoft Corporation and current Council Chairman of Hong Kong University of Science and Technology, Exploring the Low Altitude Airspace: From Natural Resource to Economic Engine (keynote video recording)
• Laurens van der Maaten, Distinguished Research Scientist at Meta AI, The Llama Herd of Models: System 1, 2, 3 Go! (keynote video recording)
• Carolina Parada, Senior Director and Head of Robotics at Google DeepMind, Gemini Robotics, Bringing AI to the Physical World (keynote video recording)

Industry Showcase: Exposition, Demos, and Sponsors

Helping to drive the future of computer vision, nearly 100 leading computer vision companies from around the globe came to CVPR 2025 to showcase the latest in commercial research and applications.

For instance, the CVPR 2025 expo delivered 69 demos, a 33 percent increase over 2024, showing an increasing focus on applied solutions. From these demos, the following were selected as this year’s top performers, based on an evaluation from organizers:

• Best Demo: DynaMem and Robot Utility Models; Haritheja Etukuru, Peiqi Liu, Chris Paxton, Soumith Chintala, Lerrel Pinto, Nur Muhammad Mahi Shafiullah
• Honorable Mention: MASt3R-SLAM; Riku Murai, Eric Dexheimer, Andrew J. Davison
• Honorable Mention: UltraFusion HDR; Yujin Wang, Zixuan Chen, Xin Cai, Zhiyuan You, Zheming Lu, Fan Zhang, Shi Guo, Tianfan Xue
• Honorable Mention: AI3D Render; Yosun Chang, Ryan Burgert

Conference platinum sponsors, who also debuted new products and research at the conference, included Adobe; Amazon Science; AMD; Apple; ByteDance; Captions; GMI; Google; Intel; Meta; Qualcomm; Sony; and Tencent. (For a complete list of sponsors, visit the CVPR site. Read exhibitor news here.)

The Intersection of AI and Art

At the intersection of art and computer vision, the CVPR AI Art Gallery featured 102 pieces from artists and researchers. Seventy-three works were featured in person at the conference and in the virtual gallery, and another 29 were featured online only.

From a review of all accepted submissions, the following three works were chosen as prize winners this year, earning a cash prize: Tom White, Atlas of Perception; Masaru Mizuochi, Green Diffusion; and Mingyong Cheng, Sophia Sun, Han Zhang, Learning to Move, Learning to Play, Learning to Animate.

Fostering Inclusivity: Broadening Participation

In addition, as sponsors of CVPR, the IEEE CS and the CVF, along with the Program Committee, committed resources to supporting broad participation from everyone in the community.

Specifically, the IEEE Computer Society Technical Community on Pattern Analysis and Machine Intelligence (TCPAMI) and CVF awarded $200,000 in travel grants and waived registrations to 276 awardees from around the globe who might otherwise have been unable to attend the conference.

The Road Ahead for Computer Vision

As new product applications continue to emerge and individuals and organizations seek more efficient, accurate, and accessible solutions, the work of the CVPR community will be instrumental in realizing the future of AI.

Watch for the latest advancements from IEEE CS and CVF throughout the year, and save the date for CVPR 2026, 6-12 June, in Denver, Colorado, US.