arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2006.07877cs.CV
收藏

FenceMask:一种预提取图像特征的数据增强方法

FenceMask: A Data Augmentation Approach for Pre-extracted Image Features

Peking University(北京大学);Beijing University of Posts and Telecommunications(北京邮电大学)

Pu Li, Xiangyang Li, Xiang Long

摘要:

我们提出了一种新的数据增强方法,名为“栅栏”,在各种计算机视觉任务中表现出出色的性能。该算法基于“模拟对象遮挡”策略,旨在实现对象遮挡与输入数据信息保持之间的平衡。通过增强遮挡块的稀疏性和规则性,我们的增强方法克服了小对象增强的困难,并显着提高性能超过基线。大量的实验证明,该方法的性能优于其他模拟物体遮挡方法。我们在CIFAR10,CIFAR100和ImageNet数据集上进行了测试,用于粗粒度分类,COCO2017和VisDrone数据集用于检测,Oxford Flowers,Cornel Leaf和Stanford Dogs数据集用于细粒度视觉分类。我们的方法在细粒度视觉分类任务和VisDrone数据集上取得了显着的性能改进。

英文摘要:

We propose a novel data augmentation method named 'FenceMask' that exhibits outstanding performance in various computer vision tasks. It is based on the 'simulation of object occlusion' strategy, which aim to achieve the balance between object occlusion and information retention of the input data. By enhancing the sparsity and regularity of the occlusion block, our augmentation method overcome the difficulty of small object augmentation and notably improve performance over baselines. Sufficient experiments prove the performance of our method is better than other simulate object occlusion approaches. We tested it on CIFAR10, CIFAR100 and ImageNet datasets for Coarse-grained classification, COCO2017 and VisDrone datasets for detection, Oxford Flowers, Cornel Leaf and Stanford Dogs datasets for Fine-Grained Visual Categorization. Our method achieved significant performance improvement on Fine-Grained Visual Categorization task and VisDrone dataset.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑