Content

This dataset represents sequential poses that can be used to distinguish 5 physical exercises. The exercises are Push-up, Pull-up, Sit-up, Jumping Jack and Squat. The dataset consists of 33 landmarks that represents several important body parts' positions. Using these landmarks, the angles and the distances between several landmarks are calculated and included in the dataset. The sequence of the poses is provided by preserving the frame order in every record.

Collection Process

About 500 videos of people doing the exercises have been used in order to collect this data. The videos are from Countix Dataset that contain the YouTube links of several human activity videos. Using a simple Python script, the videos of 5 different physical exercises are downloaded. All the frames of the videos are extracted, processed and included in the dataset.

Processing Data

For every frame, MediaPipe framework is used for applying pose estimation, which detects the human skeleton of the person in the frame. The landmark model in MediaPipe Pose predicts the location of 33 pose landmarks (see figure below). Visit Mediapipe Pose Classification page for more details. Using these landmarks, the angles and the distances between several landmarks are calculated.