Voxel51/DensePose-COCO
收藏Hugging Face2024-06-20 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/Voxel51/DensePose-COCO
下载链接
链接失效反馈官方服务:
资源简介:
DensePose-COCO是一个大规模的真实数据集,包含在COCO图像上手动注释的图像到表面对应关系。该数据集由Rıza Alp Güler, Natalia Neverova, Iasonas Kokkinos等人策划,主要用于密集人体姿态估计任务。数据集包含33929个样本,分为train和val两个分割。数据集的媒体类型为图像,包含的字段有id、filepath、tags、metadata、detections、segmentations和keypoints。数据集的创建动机和注释过程可以参考其主页和论文。
DensePose-COCO is a large-scale ground-truth dataset with image-to-surface correspondences manually annotated on COCO images. The dataset is curated by Rıza Alp Güler, Natalia Neverova, Iasonas Kokkinos, and is primarily used for dense human pose estimation. The dataset contains 33929 samples, divided into train and val splits. The media type of the dataset is image, and it includes fields such as id, filepath, tags, metadata, detections, segmentations, and keypoints. The rationale for the dataset creation and the annotation process can be referred to on its homepage and the associated paper.
提供机构:
Voxel51
原始信息汇总
数据集卡片:DensePose-COCO
数据集概述
DensePose-COCO 是一个大规模的地面实况数据集,包含在 COCO 图像上手动标注的图像到表面对应关系。
数据集详情
数据集描述
- 创建者: Rıza Alp Güler, Natalia Neverova, Iasonas Kokkinos
- 语言: 英语
- 许可证: cc-by-nc-2.0
数据集来源
- 仓库: https://github.com/facebookresearch/Densepose
- 论文: https://arxiv.org/abs/1802.00434
- 主页: http://densepose.org/
用途
密集人体姿态估计
数据集结构
- 名称: DensePoseCOCO
- 媒体类型: 图像
- 样本数量: 33929
- 持久性: False
- 标签: []
- 样本字段:
- id: fiftyone.core.fields.ObjectIdField
- filepath: fiftyone.core.fields.StringField
- tags: fiftyone.core.fields.ListField(fiftyone.core.fields.StringField)
- metadata: fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.metadata.ImageMetadata)
- detections: fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.labels.Detections)
- segmentations: fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.labels.Detections)
- keypoints: fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.labels.Keypoints)
数据集包含两个拆分:“train”和“val”。样本根据其拆分进行标记。
数据集创建
创建动机
请参考主页和论文以获取创建动机。
标注过程
请参考 GitHub 仓库以获取标注过程。
引用
BibTeX: bibtex @InProceedings{Guler2018DensePose, title={DensePose: Dense Human Pose Estimation In The Wild}, author={Rıza Alp Güler, Natalia Neverova, Iasonas Kokkinos}, journal={The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}, year={2018} }



