Last week, we announced the General Availability of the Rendered.ai synthetic data platform. If you’re wondering “What is a synthetic data platform?” head to our website. If you’ve done that and are now wondering “OK, but how does it work?” Wonder no longer! We’ve put together a video that shows you just what we mean by a PaaS for synthetic data.
In just over 8 minutes (or 4 minutes at double speed, let’s be real), I walk you through the end-to-end process of signing into your organization, creating a workspace, editing a simulation process chain (known as a Graph), and leveraging our managed cloud compute backend to generate a dataset with full annotations and metadata for use in training a computer vision model.
In the video, you’ll see a dataset generated from our example channel showing toys in a toy box, but the possibilities with physics-based simulation are endless. Want to train a detection model to find rare objects in satellite imagery? You can do just that. Do you need the ability to detect specific hazardous items under X-ray? You can do that too. Need to train a model on Synthetic Aperture Radar (SAR) returns from a sensor that hasn’t even been manufactured yet? Well, no one has done that yet, but you could be the first!
This video just scratches the surface of what is possible with the Rendered.ai platform, but it’s all you really need to take the most important step in your synthetic data journey: getting started.