Welcome to the repository of the COCO Single Person Keypoints Dataset. This is a processed subset of the famous COCO Dataset, where each image contains exactly one person. The dataset is particularly suitable for tasks related to human pose estimation.

The dataset is created by processing the original COCO 2017 dataset. It extracts only those images where there's exactly one person in the image. The images are cropped to the bounding box of the person, and the annotations are updated accordingly.

The dataset follows the same structure as the original COCO dataset. There are three main types of data:

• Images: Located in the data/coco_single_person/val2017 directory.
• Annotations: The annotations file is found at data/coco_single_person/annotations/person_keypoints_val2017.json.
• Image info: Additional information about the images can be found in the annotations file.

The dataset can be used in a similar way as the COCO dataset. For Python users, we recommend the pycocotools library which provides API to read and manipulate the data. Check out this notebook for a quick introduction to using pycocotools with a COCO styled dataset.

You can evaluate your pose estimation models on this dataset using the COCO Evaluation metric. An example code for this is given in this repository.

This repository is open to contributions. If you encounter any issues while using the dataset or have suggestions for improvements, please open an issue. For major changes, please open an issue first to discuss what you would like to change.