<ul>
<li>1,000 videos (with 5-10 exercise reps each)</li>
<li>10 exercises</li>
<li>7 unique indoor scenes</li>
<li>Realistic environmental occlusion</li>
<li>Diverse lighting conditions</li>
<li>Varied body shape, skin tones, and clothing</li>
<li>Rich annotations for 2D and 3D supervision</li>
<li>224x224 RGB videos (24fps)</li>
</ul>

The dataset includes 10 of the most common fitness and physical therapy (PT) exercises.
<ol>
<li>Pushups</li>
<li>Alternating Bicep Curls (with dumbbells)</li>
<li>Delt Flys (with dumbbells)</li>
<li>Squats</li>
<li>Bird Dogs</li>
<li>Supermans</li>
<li>Bicycle Crunches</li>
<li>Leg Raises</li>
<li>Front Raises (with dumbbells)</li>
<li>Overhead Press (with dumbbells)</li>
</ol>

InfiniteRep videos are accompanied by a rich set of pixel-perfect labels and annotations, including:
<ul>
<li>Frame-specific rep counts</li>
<li>Avatar characteristics (gender, clothing, waist circumference, etc.)</li>
<li>Camera pitch and height</li>
<li>Skeleton keypoints (17-point COCO skeleton plus XXXX)</li>
<li>Bounding boxes</li>
<li>Segmentation masks</li>
<li>Occlusion metadata</li>
<li>Much more!</li>
</ul>
For the full description of labels and metadata, check out the <a href="https://github.com/toinfinityai/infinity-datasets/tree/main/fitness-basic" target="_blank">README</a>.

224x224 RGB videos (24fps)
<br>
2.5 GB total dataset
<br>
Each of the 10 exercises is in a separate zipped folder (400-800MB in size)

This dataset is licensed under a <a href="https://creativecommons.org/licenses/by/4.0/" target="_blank">Creative Commons Attribution 4.0 International License</a>.
<br>
Both academic and commercial applications are allowed.

A newer version of this dataset is available for free (InfiniteRep 2.0). It is similar to InfiniteRep except with higher-resolution videos (512x512 px) and additional annotations (like 3D keypoints and camera intrinsics/extrinsics).

<ul>
<li>Github <a href="https://github.com/toinfinityai/infinity-datasets/tree/main/fitness-basic" target="_blank">README</a>: full dataset and annotation descriptions</li>
<li>Demo <a href="https://github.com/toinfinityai/infinity-datasets/blob/main/fitness-basic/quickstart.ipynb" target="_blank">Jupyter notebook</a>: how to filter the dataset based on metadata and visualize labels</li>
</ul>

Questions? We’re happy to chat asynchronously via email or hop on a call. Just send us a note at info@infinity.ai (this goes to all of the Infinity AI founders).