You Only Cut Once: Boosting Data Augmentation with a Single Cut
This is an incremental improvement for computer vision practitioners, offering a parameter-free way to enhance data augmentation diversity and model robustness.
The paper tackles the problem of limited diversity in data augmentation by introducing YOCO, which cuts an image into two pieces and augments each piece separately, improving object recognition from partial information. The method boosts performance on CIFAR and ImageNet classification tasks, sometimes surpassing conventional augmentations by large margins, and enhances contrastive pre-training for better transfer to downstream tasks.
We present You Only Cut Once (YOCO) for performing data augmentations. YOCO cuts one image into two pieces and performs data augmentations individually within each piece. Applying YOCO improves the diversity of the augmentation per sample and encourages neural networks to recognize objects from partial information. YOCO enjoys the properties of parameter-free, easy usage, and boosting almost all augmentations for free. Thorough experiments are conducted to evaluate its effectiveness. We first demonstrate that YOCO can be seamlessly applied to varying data augmentations, neural network architectures, and brings performance gains on CIFAR and ImageNet classification tasks, sometimes surpassing conventional image-level augmentation by large margins. Moreover, we show YOCO benefits contrastive pre-training toward a more powerful representation that can be better transferred to multiple downstream tasks. Finally, we study a number of variants of YOCO and empirically analyze the performance for respective settings. Code is available at GitHub.