How Synthetic Data Presents Opportunities for Airbus Customers

Airbus is producing 100-times more data than it did 10 years ago. The only way its customers can interpret this data is to use machine learning (ML) to automatically extract insights from the imagery. To achieve this they need good, large and diverse training datasets. However, access to labeled data is holding them back from training those algorithms. OneView and Airbus joined forces to prove that synthetic data is a valid replacement for real imagery to train ML algorithms.


During this presentation you will discover how the results of this joint project validates the use of synthetic data to train algorithms, solving this training data bottleneck by:


1. Replacing the manual collection and annotation of data

2. Increasing model’s accuracy

3. Dramatically cutting the costs of data collection and annotation

4. Reducing data sourcing time

5. Training models where real-world data is not available.