Data Warehouse Designed for Synthetic Data

Mirage ingests, organizes, and analyzes all formats of perception data to empower machine learning teams with a consolidated mix of real and synthetic data.

Early Access
Sensor realistic drone footage of mountaintops

why mirage

What are current problems?

01

Filling perception dataset gaps is hard because real data is expensive, hard to gather, and incomplete.
--->

The average cost of real data is $1 per labelled image, while 80% of it goes unused. In addition, it is hard to gather, with an average of 2 month turnaround, and incomplete (edge cases).

02

There is no solution that is able to handle and analyze every data format 2D (png, jpeg), 3D assets (USD, FBX), pointclouds (Numpy), etc and use data augmentation to optimally improve ML models.

Perception data storage often suffers from an inability to handle all data formats, especially with synthetic data.
--->

What's the solution?

Mirage’s platform consolidates all of the customer’s perception data (real and synthetic). Mirage handles and analyzes every data format and uses active learning to find dataset gaps.

Our platform empowers engineers to improve perception datasets through numerous methods including: adding synthetic assets to real images, domain randomization, adapting to a new sensor, creating synthetic 3D environments from real images, removing the sim-to-real gap with sensor calibration, etc. 

Ready to explore Mirage?

Reach out!

Early Access

Used at companies you know.

A green banner to call out an exciting new feature or quote from a customer