Athanasios Angelakis

Data Centric AI: How to “Not” Use Data Augmentation

Deep Learning has tremendous applications in different fields of our every day life. Depending on the field, domain experts provide their input to a data science pipeline such that a deep learning model will, somehow, take advantage of this knowledge. Unfortunately, many of these domain experts or even deep learning engineers, understand in a wrong way some fundamental definitions of computer vision. Using the hardest modality of radiology, namely, Ultrasound we will try to concretely understand how to use data augmentation in a data centric approach in order to create more robust and ready to production deep learning models.

This website uses cookies to ensure you get the best experience. Learn more