Skip to main content
. 2022 Apr 1;2022:4727375. doi: 10.1155/2022/4727375

Table 2.

Human pose estimate datasets.

Dataset Feature
LSP The data comes from sports category labels, image zoom, only annotated in each picture-one person, 14 key points
FLIC Data comes from hollywood movies, 10 key points
MPII human pose The data comes from YouTube videos, which covers 410 human activities, each image has an activity tag, and 16 key points
MSCOCO The data comes from the internet, it contains various activities, 17 key points
Al challenge The data is captured from the internet. It is currently the largest human pose image data set, and 14 key point
PoseTrack The data comes from the MPII human pose data set and 15 key points
Buffy The data comes from a TV show. Line segments are provided to indicate the location, and the size and direction of the body parts are also provided