Handshape Library

A python library to analyze and annotate handshape clusters. The tool receives as inputs the extracted json files from Openpose and lets the user run k means on the normalized and scaled hand configurations.

Getting started

Prerequisites

You need to install the following python libraries in order to use the library

os,
PIL,
cv2,
natsort,
pandas,
matplotlib,
sklearn,
numpy,
bokeh,
moviepy

Running your own data

To run the framework with your own data, you simply need to run Openpose in the video you want to analyze. Get the json files and the output video and place them in a folder along with this library. See Example.ipynb for the supported functions. If you wish run:

from HSL import Annotations
Annotations.annotate("json_directory/", "frame_directory/").video

to get everything with the predifined values