Edge Impulse’s Time Series Data Augmentation has been shortlisted for the Electronics Excellence Awards under the Software category.
It’s expensive and time consuming to collect enough data to train production-ready ML models. Time Series Data Augmentation is an Edge Impulse feature that programmatically increases the size and diversity of time-series training datasets by applying controlled, statistically grounded signal transformations. It is designed for sensor-based ML workloads where labelled data is often limited, expensive to capture, or highly variable in the real world.
The block supports four augmentation modes to create realistic synthetic samples that help models generalise better, resist overfitting, and remain robust under real-world noise and variation.
The primary users are machine learning engineers, embedded developers, and data scientists working with time-series signals (such as IMU, audio, and vibration data) who need to improve model accuracy and resilience using richer training sets without collecting additional real-world data.
For more information, visit: https://www.edgeimpulse.com/blog/betting-on-decomposition-based-time-series-augmentation/
The Electronics Excellence Awards ceremony will take place at embedded world on 11th March at 4.30pm at the Hall 5 Forum.
You’re all welcome to come and see which product will win!
