This project contains my Rose-Hulman undergraduate thesis on mobile hand pose estimation with deep learning.
In order to run this project, you will need the following packages:
There are several ways to install these, namely Pip and Anaconda; I recommend using Anaconda.
If you're impatient, you can simply run the following in a Linux terminal to get up and running:
git clone [email protected]:zanecodes/thesis.git
cd thesis
wget http://horatio.cs.nyu.edu/mit/tompson/nyu_hand_dataset_v2.zip
unzip nyu_hand_dataset_v2.zip -d nyu_hand_dataset_v2
pip3 install -r requirements.txt
jupyter notebook
The NYU Hand Pose Dataset is 92gb, so be sure you have the disk space and bandwidth for this, and expect it to take a while. You should now have a web browser open with a running instance of Jupyter Notebook.