iMet Collection 2019
收藏资源简介:
iMet Collection 2019数据集是由谷歌研究和大都会艺术博物馆合作创建的,专注于细粒度的艺术品属性识别。该数据集包含155,531个样本,每个样本都附有由世界级博物馆专家审核的高质量属性标签。数据集的内容涵盖了多种艺术品,包括绘画、乐器等,数据来源于大都会艺术博物馆的开放访问项目。创建过程中,数据集的属性标签由专家和外包组织共同标注,确保了标签的准确性和多样性。该数据集主要应用于计算机视觉领域,旨在推动细粒度视觉识别技术的发展,特别是在艺术品数字化管理方面。
The iMet Collection 2019 Dataset was developed through a collaboration between Google Research and the Metropolitan Museum of Art, focusing on fine-grained artwork attribute recognition. It contains 155,531 samples, each accompanied by high-quality attribute labels reviewed by world-class museum experts. Covering diverse artwork categories including paintings, musical instruments and more, the dataset is sourced from the Metropolitan Museum of Art's Open Access program. During its construction, the attribute labels were jointly annotated by experts and outsourcing institutions, which guarantees the accuracy and diversity of the labels. This dataset is primarily utilized in the field of computer vision, with the goal of promoting the advancement of fine-grained visual recognition technologies, especially in the domain of artwork digital management.

- 1The iMet Collection 2019 Challenge Dataset谷歌研究 · 2019年



