数据集名称未明确提供
收藏资源简介:
本研究涉及的数据集涵盖了多个领域,包括自然图像、医学影像、卫星数据等,共计六个数据集。这些数据集的创建旨在评估数据效率高的图像分类方法,特别是在数据量有限的情况下。每个数据集都经过子采样以适应小数据场景,确保每个类别有大约50个训练图像。数据集的应用领域广泛,从日常物体的自然图像分类到专业领域的医学影像分析,旨在解决在数据稀缺情况下如何有效训练深度神经网络的问题。
The datasets utilized in this study span multiple domains, including natural images, medical imaging, satellite data and others, totaling six datasets. These datasets are developed to evaluate data-efficient image classification approaches, particularly in scenarios with limited training data. Each dataset has been subsampled to adapt to small-data settings, ensuring approximately 50 training images per category. The datasets cover a broad range of application fields, from natural image classification for daily objects to medical image analysis in professional domains, aiming to address the issue of how to effectively train deep neural networks under data scarcity.




