Skip to content

Voxel51 at CVPR 2024!

CVPR 2024 is next week! Here’s when and where you can find members of the Voxel51 team. We encourage you to put a visit to the Voxel51 team on your CVPR 2024 agenda – we’d love to meet you!

But first, you’re in for a special treat this week. It’s the CVPR Preshow! Hacker-in-Residence Harpreet Sahota and Chief Scientist Jason Corso interviewed authors who presented research in vision-language models, 3D computer vision, and diffusion models. Watch these interviews to get a deeper insight into what type of bleeding-edge research is being presented at CVPR 2024 Follow Harpreet on LinkedIn to be among the first to see preshow interviews as they are posted.

Visit Voxel51 booth #1519 to …

See the mind-blowing open source FiftyOne project in action

If you’re new to Voxel51, we’re developers of the open source FiftyOne project that helps visual AI builders improve data quality and boost model performance. We love showing the mind-blowing workflows made easy with FiftyOne. Stop by booth #1519 anytime and see, for example, how to use FiftyOne to:

  • Easily curate and manage your data
  • Evaluate models
  • Visualize embeddings
  • Find annotation mistakes
  • Find image quality issues
  • Reverse image search
  • Zero-shot prediction
  • Mine hard samples
  • Work with multi-modal data (3D, images, videos, more!)

Or, tell us what you’re working on and we’ll show you how FiftyOne can help!

If you’ve already visited us at CVPR, no fear — there are lots of new goodies we think you’ll like. In fact, since last year’s CVPR, there have been 20 new releases of open source FiftyOne, including the latest 0.24.0 release with 3D meshes, custom workspaces, and more!

Score some coveted swag

We’re confident you’ll walk away knowing how FiftyOne can instantly alleviate some of the pain points you’re facing in your visual AI projects, but we’re also giving away some sweet swag that can be yours while supplies last.

Learn about our open roles

Voxel51 is growing quickly and we are looking for people to grow with us. Visit our recruiting desk in our booth to learn more about our open roles (below) and meet members of the team to find out what it’s like working here.

Say hi and let us know what you’re working on!

We’re AI/ML builders and enthusiasts, too, and we love hearing what fellow members of the community are working on. Stop by and let us know what you’re up to – your latest research, AI apps you’re building, datasets you’re building, models you’re fine-tuning, or whatever’s on your mind. 

Check out these two workshops with Voxel51 Chief Scientist Jason Corso

Jason Corso, Professor of Robotics, Electrical Engineering, and Computer Science at the University of Michigan and Co-Founder / Chief Science Officer of Voxel51, will speak in two workshops this year at CVPR. Have a look at the details and we hope to see you there!

Computer Vision with Humans in the Loop (CVHL)

  • Date: Tuesday, June 18, 2024
  • Time: 8:30 AM5:45 PM
  • Room: Room Summit 329
  • More Details:

Summary: This workshop aims to explore the pivotal role of human interaction in advancing computer vision technologies. Despite significant progress over the past two decades, computer vision systems often fall short of human capabilities, especially in specialized or complex scenarios. Featuring a series of invited talks and panels, this workshop will highlight innovations like interactive segmentation, advancements in language and visual prompt integration, and the development of self-aware systems capable of recognizing and compensating for their limitations. Distinguished speakers from both academia and industry will share their insights, offering a rich dialogue on the integration of human cognitive skills with machine precision to tackle the challenges of computer vision. Participants will gain an understanding of the evolving landscape of human-in-the-loop methodologies and their impact on practical applications and foundational models in the field. Join us to discuss historical perspectives, current research, and future directions in this dynamic area.

Visual Odometry and Computer Vision Applications Based on Location Clues

Summary: Visual odometry and localization have maintained an increasing interest in recent years, especially with the extensive applications for autonomous driving, augmented reality, and mobile computing. With the location information obtained through odometry, services based on location clues are also rapidly emerging. Particularly, in this workshop, we focus on mobile and robot platform applications. This workshop invites papers in the areas including advances in visual odometry and computer vision applications based on location context.

Learn about embeddings and 3D in a flash

We signed up to present two Flash Sessions! Flash sessions are 15-minute TED-style talks where you can learn a lot about computer vision techniques in a short amount of time. Check out our two sessions, in the Expo Hall on stage at booth 1841.

Thursday, June 20, 3:00-3:15 pm: 

5 Handy Ways to Use Embeddings, the Swiss Army Knife of AI

Presented by Hacker-in-Residence Harpreet Sahota

Discover the incredible potential of vector search engines beyond RAG for large language models! Explore 5 handy embeddings applications: robust OCR document search, cross-modal retrieval, probing perceptual similarity, comparing model representations, concept interpolation, and a bonus—concept space traversal. Sharpen your data understanding and interaction with embeddings and open source FiftyOne.

Friday, June 21, 12:00-12:15 pm:

Build Your Own Virtual World: 3D Reconstruction in FiftyOne 

Presented by ML Engineer Daniel Gural

3D is one of the fastest-growing spaces in ML, and new models are coming out that can achieve incredible results. In this talk, you’ll learn some of the methods used to create 3D reconstructions, the drawbacks of today’s models, and what there is to be excited about on the horizon.