HOW TO USE THIS PROJECT

A Review of Deep Learning Techniques for Markerless Human Motion on Synthetic Datasets

Based on DeepLabCut toolbox, we run with three synthetic datasets

• Easy Pose Datasets: contains T-pose, A-pose, standing and seating
• Inter Pose Datasets: has walking and running
• Hard Pose Datasets: includes all the postures with high complexity like yoga, push-ups, and activities

Our project is available at https://github.com/DoanDuyVo/DeepLab_Human

Research paper is also available at https://github.com/DoanDuyVo/DeepLab_Human/blob/main/DeepLab_Human_Paper.pdf

The original version of DeepLabCut can be found step-by-step in the user guide at https://github.com/DeepLabCut/DeepLabCut

Here are the results of project:

Figure 1: Images from evaluation results

Figure 2: The graphs plot the trajectories

Click on the images to watch video which are created with trailpoints and draw_skeleton

More resources for help

If you are new to DeepLabCut, here are resources we recommend before jumping in.

• Please first read at the Nature Protocol paper;
• Check out the quick video on navigating the docs;
• Check out the free DeepLabCut Course: we have put together a "course" on the science of DeepLabCut and how to use it.

There are the links to all the key steps to get you up and running within a day

• How to install DeepLabCut
• How to use standard DeepLabCut
• How to use multi-animal DeepLabCut

References:

• If you use this code or data provided by DeepLabCut, we kindly as that you please cite Mathis et al, 2018 and, if you use the Python package (DeepLabCut2.x) please also cite Nath, Mathis et al, 2019.

• If you use our dataset, images or run code provided by this project, please cite Doan.V & Butler.T, 2021