The AI boom is fueled by two things: computing power and data. Real-world data can be costly and time-consuming to collect, creating a bottleneck in the process. Synthetic data can alleviate this bottleneck by providing scalable diverse datasets with perfect annotations and labels.
Unreal Engine, using only native tools and features, can produce synthetic visual data for virtually any use case. Synthetic data is currently being used for computer vision applications at companies like Boeing, Tesla, Microsoft, and Apple to train models for a fraction of the cost of manually collecting it. Synthetic data also allows for edge cases that would be too costly or dangerous to enact in the real world.
To aid in this endeavor, Boeing created custom tools in Unreal Engine using Editor Utility Widgets, enabling data scientists with very little knowledge of Unreal Engine to customize and generate their own data sets.
In this session recorded at Unreal Fest Orlando 2025, Taylor Loper of The Boeing Company walks you through how to use Editor Utility Widgets to ingest data (flight logs, car paths, CAD properties) and use that data to populate Sequencer; describes the process of labeling or annotating images generated by Unreal Engine in various formats to make them compatible with common architectures and frameworks like YOLOv8; and discusses the Unreal Engine’s ability to emulate camera intrinsic properties.
Learn more about Unreal Engine here: unrealengine.com
#Boeing, #AI, #SyntheticData, #EditorUtilityWidgets, #Sequencer, #UnrealEngine, #UE5, #UnrealFestOrlando