1 paper · 1 filter
Suorong Yang, Jie Zong, Lihang Wang +6
Data augmentation has been widely employed to improve the generalization of deep neural networks. Most existing methods apply fixed or random transformations. However, we find that…