better vision Models
with SBX Robotics

Synthetic data is an exciting new way to accelerate computer vision. Training with synthetic data produces robust deep learning models, using only a small fraction of real data. Using synthetic data is the right way to build and improve computer vision models

Companies who choose us

Streamlined Data Processing

Training on synthetic data helps you develop powerful deep learning models without having to collect, label, and clean real data. This is currently the fastest way of bootstrapping and improving computer vision models.

Get Time & Cost-Effective Training

Using simulation software enables us to create training data that’s up to 10x faster and more cost-effective than other annotation services or in-house teams.

Optimized for Top Performance

Every SBX dataset goes through rigorous testing and optimization to ensure the best possible performance on real-world data.

The Power of SBX Robotics

SBX Robotics uses technology from film and gaming to create realistic, perfectly labeled training datasets that support object detection, segmentation and 6D post estimation models.

  • Clients send 25 images from their robot’s camera and in turn, receive 25,000 perfectly labeled synthetic training images. SBX’s data is ready for use by deep learning computer vision models.
  • Synthetic data gives data scientists and developers full control. Long gone are the days of unreliable, incomplete data. You don’t have to struggle to find data for machine learning at any scale.
  • Real data tends to be supplemented with synthetic data as a means of creating enhanced observations and trends.

Case Study

Detecting forklifts with OTTO Motors

The team at OTTO Motors looked to SBX synthetic data to bootstrap and improve computer vision systems in their line of AMRs.

Otto Motors & SBX Robotics collaboration video thumbnail

What Customers say

John Novak
Head of Computer Vision

"With SBX synthetic data, we hope to never use real data again!"

Dave Weatherwax
Senior Director of Software

“SBX generates high fidelity synthetic data with a rapid turnaround time”

James Servos
Perception Team Manager

"SBX kickstarts projects far faster than real-world data acquisition and labeling."

Daniel Grollman
Lead R&D Engineer, PhD

"SBX showed us synth data is viable for our AI needs. We look forward to working together."

Marek Cygan
CTO

"The ability of models trained on SBX data to generalize on fairly diverse items is impressive."

Tarik Kelestemur
Researcher

"Our model trained on SBX data significantly outperformed one trained on data we collected."

Try out our platform, it’s easy!

Share 25 images from your vision system, and we will generate an optimized training set of 25,000 annotated synthetic images.