Data augmentation is a technique used in machine learning and computer vision to increase the amount of data available for training models. It involves creating new and varied versions of existing data by applying transformations such as rotations, flips, changes in brightness or contrast, and adding noise or distortions. These modified input images can help improve a model's accuracy and performance by making it more robust to variability and real-world conditions. In essence, data augmentation expands the diversity of a dataset, and as such, helps to address issues related to overfitting, and underfitting.
Publication date: