Now that we have an overview of human pose estimation and the new DeeperCut algorithm, we can run the code for single-person pose detection and check that out in the Jupyter Notebook.
We will start with single-person detection. Before starting, we need to make sure that we are using a clean kernel. You can restart your kernel, or you can use the hotkeys to do the same. You can then press the 0 key twice when you're in command mode, which is opposed to edit mode when you're actually editing the cells.
Let's start with our single-person detection code, as shown in the following example:
!pip install pyyaml easydict munkres
The exclamation mark means execute a shell command. This will install a couple of libraries that you might not have, and if you have Python 3 installed in your system, you might need to change the command to pip 3.
...
