Simplifying Kinect for Windows 1.x Skeleton Drawing

KinectTools is an abstraction over the code that handles the Kinect v1 sensors, Skeleton data, and drawing the skeleton. It exposes Brush and Pen properties for drawing so you have control over the look of your skeleton. It also can place a PNG file over the head as you move around, providing hours of jocularity.

If you've done any work with the Kinect for Windows 1.x SDK you've probably already created an abstraction such as this. But if you haven't here's a nice one for you.

What's cool about this is that it uses the term Body which is what SDK 2.0 calls a Skeleton. I've also written this abstraction for SDK 2.0 (only in pre-release) so using this will get you ready for the future. The next version of the GesturePak Recorder and sample code uses this abstraction as well.

Here's a very simple WPF app that uses the KinectTools SimpleBodyEngine class to draw the skeleton in real time, put an image of my head on top of it, and turn the background of the window red if you move your right hand out in front of your body 1/3 of a meter.

XAML:

<Window x:Class="SimpleBodyEngineTest.MainWindow"
    xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
        xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
        Title="MainWindow" Height="600" Width="800">
    <StackPanel>
        <StackPanel Orientation="Horizontal">
            <TextBlock Text="Sensor: "/>
            <TextBlock Text="{Binding SensorStatusName}" />
        </StackPanel>
        <Image x:Name="BodyImage" Source="{Binding ImageSource}" Stretch="None" />
    </StackPanel>
</Window>

Code:

using System;
using System.Windows;
using System.Windows.Media;
using Microsoft.Kinect;
using KinectTools;

namespace SimpleBodyEngineTest
{
    /// <summary>
    /// Interaction logic for MainWindow.xaml
    /// </summary>
    public partial class MainWindow : Window
    {
        // This object makes handling Kinect Skeleton objects easy!
        SimpleBodyEngine BodyEngine = new SimpleBodyEngine();

        public MainWindow()
        {
            InitializeComponent();
            // event handlers
            this.Closing += MainWindow_Closing;
            BodyEngine.BodyTracked += BodyEngine_BodyTracked;

            // put Carl's head on top of the Skeleton
            BodyEngine.HeadImageUri = new Uri("carl.png", UriKind.Relative);
            
            // bind the XAML to the SimpleBodyEngine
            this.DataContext = BodyEngine;
        }

        /// <summary>
        /// This event fires when a *tracked* skeleton frame is available
        /// </summary>
        /// <param name="sender"></param>
        /// <param name="e"></param>
        void BodyEngine_BodyTracked(object sender, BodyTrackedEventArgs e)
        {
            // Get the Z position of the hand and spine
            var hand = e.Body.Joints[JointType.HandRight].Position.Z;
            var spine = e.Body.Joints[JointType.Spine].Position.Z;

            // if the hand is in front of the spine...
            if (Math.Abs(hand - spine) > .3)
                // turn the background red
                Background = Brushes.Red;
            else
                Background = Brushes.White;
        }

        void MainWindow_Closing(object sender, System.ComponentModel.CancelEventArgs e)
        {
            BodyEngine.Dispose();
        }
       
    }
}

Forget about the hundreds of lines of code to draw the Skeleton. If you just want to handle the data, read this blog post I wrote on the basics of Skeleton tracking. This code is so simple. Put up an image and bind it to a property. Create a new SimpleBodyEngine object and make it the DataContext. Done.

Download the code here and enjoy.

Carl

Comments

[Ottavio Nuccio - Abito sposo]

Thank you Carl for sharing this useful information! :)