Synthetic Data

Our Synthetic Data Machine Learning platform is powered by physical based rendering and procedural shader framework. It is able to simulate various scenarios, lighting and weather conditions. Use it to accelerate the process of machine learning and generate high-confidence datasets.

Crisis Scenarios

Our platform generates learning models where traditional data collection isn’t possible. We are able to generate scenarios such as fire drill, car accident, road crack, etc, for computer vision A.I. system, removing the need of real world data capturing.

Procedural Shader Generation

We implement our own shader framework to simulate different surface properties of the object, including terrain and weather. Our procedural generation system allows us to adjust and programmatically generate thousands of different object appearances from old to new, dry to wet, good to damage.

Auto Annotation

The systems automatically locate the parts of the object with the most characteristics and annotate them. Users can also define specific selected parts to annotate in the system and generate data in desired format for machine learning.

Dynamic Camera

Different camera angles and filters can be adjusted according to users' requirements. We support night vision, infrared, heat sensor and any type of camera filters in 360 degree without the need of a real camera present.