可解释抽象火车数据集 (XTRAINS)

可解释抽象火车数据集（XTRAINS）是由新里斯本大学科学与技术学院创建的一个图像数据集，包含500,000张152×152像素的简化火车图像。该数据集旨在为算法的研究和应用提供平台，特别是关于解释和提取解释的研究。数据集中的每张图像都通过一个详细的分类学进行标注，描述了火车的视觉特征，并提供了边界框以精确标注火车元素。数据集的创建过程涉及随机生成火车的各种特征，如车轮大小、连接器高度等，以确保多样性和复杂性。该数据集主要应用于可解释人工智能领域，旨在解决如何从人工智能模型中提取和理解解释的问题。

The eXplainable Train Abstract Image Dataset (XTRAINS) is an image dataset developed by the School of Science and Technology of NOVA University Lisbon, containing 500,000 simplified train images with a resolution of 152×152 pixels. This dataset aims to provide a platform for algorithm research and application, particularly studies focused on interpreting and extracting explanations. Each image in the dataset is annotated using a detailed taxonomy that describes the visual features of the train, and is accompanied by bounding boxes for precise annotation of train components. The dataset creation process involves randomly generating various train features such as wheel size and coupler height to ensure diversity and complexity. This dataset is primarily utilized in the field of Explainable Artificial Intelligence (XAI), targeting the problem of how to extract and comprehend explanations from artificial intelligence models.