Overfitting(過学習)を防ぐ → ⽬的に合わないパターンや、実際のデータに出てこない バリエーションは避けること Ø Augmentationの例 ü Horizontal Flip ü Vertical Flip ü Random crop ü Rotation ü Channel shift ü Width shift range ü Height shift range ü Zoom
DeVries and G. W. Taylor, "Improved Regularization of Convolutional Neural Networks with Cutout," in arXiv:1708.04552, 2017. • Z. Zhong, L. Zheng, G. Kang, S. Li, and Y. Yang, "Random Erasing Data Augmentation," arXiv:1708.04896, 2017.