This repo contains a collection of tools for assessing the performance of pose estimation algorithms, in both single-view and multi-view setups.

Installation

First you'll have to install the git package in order to access the code on github. Follow the directions here for your specific OS. Then, in the command line, navigate to where you'd like to install the package and move into that directory:

$: git clone https://github.com/paninski-lab/tracking-diagnostics
$: cd tracking-diagnostics

Running in an existing conda environment

The dependencies for this package are relatively minimal, and quite standard - numpy, scipy, matplotlib, etc. It may be possible to run this code in an already existing environment, with few additional package installations. Please stay tuned for more updates.

To make the package modules visible to the python interpreter, locally run pip install from inside the main tracking-diagnostics directory after you have activated an existing conda environment:

(existing_environment) $: pip install -e .

Creating a new conda environment

To create a new environment specifically for this package, follow the directions here to install the conda package for managing development environments. Then, create a conda environment:

$: conda create --name=diagnostics python=3.9
$: conda activate diagnostics
(diagnostics) $: pip install -r requirements.txt

To make the package modules visible to the python interpreter, locally run pip install from inside the main tracking-diagnostics directory:

(diagnostics) $: pip install -e .

To be able to use this environment for jupyter notebooks:

(diagnostics) $: python -m ipykernel install --user --name diagnostics