Kinect – How It Works
Kinect, the upcoming gaming technology being developed by Microsoft, is using sophisticated video and sensor technology combined with cutting edge advances in artificial intelligence to develop a new gaming system that doesn’t require the use of controllers to play games. Instead, players uses body movement and voice commands to control the game. This frees gamers from having to learn how to use controllers and other devices to play videos games which currently act as barriers for some people. The new technology also means that players not only control the games with their own bodies but become part of the game itself in a 3-D environment.
So how does Kinect work? Kinect is an add-on that works with Microsoft’s Xbox 360. The hardware consists of a sensor that contains two cameras, along with a multi-array microphone. The sensor is encased in a box that sits either on or near your television set and connects with your XBox 360.
Inside the sensor is an RGB camera with facial recognition and fully body tracking capabilities. The other camera is a depth sensor that judges distances and sees a room in 3-D. The multi-array microphone enables the system to locate up to four voices in a room at one time. Taken together, what this means is that the gaming system can see and track up to four players in a room as well as recognize their faces, voices and full body movements as they play a game, allowing players to experience games in 3-D and eliminating the need for controllers.
However, the brains behind the hardware is what really makes this all possible. This comes from the custom processors and proprietary software that Microsoft is developing for Kinect. It’s one thing for the cameras to see what is going on inside a room. It’s quite another to teach the cameras to understand what they see.
In order for the gaming system to react to players’ movements in a game, the system must recognize individual faces, voices, body parts and body movements. In order to do this Microsoft is integrating sophisticated advances in artificial intelligence, particularly in the field called machine learning, to teach the Kinect system to make sense of all the images that it captures with its cameras.
Machine learning for Kinect involves feeding a network of computers with millions of different images of people in various positions so that the system can learn and be able to understand and interpret what it is seeing. This saves on having to hand code all of the millions upon millions of potential body movements that humans can make.
What this allows Kinect to do is recognize specific body parts, transform them into a simple avatar, and repeat the process 30 times per second as players continue to move and react during a game. The process happens fast enough for the gaming system to take all the information in and use it to control the game. This sophisticated software is what allows you to control the game with your body movement and voice instead of having to use a controller.
Kinect’s release date is set for November 4, 2010. If you want more information about Kinect, including the price and how to pre order, as well as finding out all of the games that will be compatible with Kinect, then sign up for our newsletter below, and be in the know!
Tagged with: No available tags
