Apple to sponsor and present new research at annual computer vision conference

Apple is headed to Nashville next week, where it will showcase new computer vision research and run live technical demos at this year’s IEEE/CVF Conference on Computer Vision and Pattern Recognition.

Three papers, one demo

This year’s CVPR will take place June 11–15 in Nashville, Tennessee, and Apple will be on-site presenting three recent white papers it published, two of which we covered recently at 9to5Mac:

• FastVLM: Efficient Vision Encoding for Vision Language Models, a model that dramatically speeds up real-time high-res image understanding by using fewer visual tokens
• Matrix3D: Large Photogrammetry Model All-in-One, a unified model that simplifies 3D content creation even when training data is incomplete.
• World-Consistent Video Diffusion with Explicit 3D Modeling, a model that boosts 3D consistency and enables more accurate 3D structure prediction when the camera position is unknown.

The IEEE/CVF CVPR has been held since 1983, and is one of the most competitive and influential events in the field. Fewer than 5% of submissions are selected for oral presentations.

Apple will be demoing FastVLM on these days and times:

• Friday, June 13: 10:00 am – 12:30 pm, 2:30 pm – 4:30 pm
• Saturday, June 14: 10:00 am – 12:30 pm, 2:30 pm – 4:30 pm
• Sunday, June 15: 10:00 am – 12:30 pm

Also, more than 20 Apple-affiliated researchers are listed as reviewers for the conference, a sign of just how deeply the company is looking to integrate with the academic side of the AI ecosystem.