Kinect’s Evolution: Then versus Now
- Nov 9, 2011
First introduced to the world as “Project Natal” on June 1, 2009 at E3, Kinect was launched in the United States on November 4, 2010. In just 60 days, more than 8 million units were sold worldwide; and by March 2011, more than 10 million units were in consumers’ hands.
Kinect continues to sell well, which is not surprising considering Microsoft’s continued investment in improving Kinect’s tracking sensitivity, voice recognition, and integration into the Xbox 360 dashboard. Most of the limitations that existed at launch have been overcome, either by software updates or the development of products by third-party vendors. In addition, Kinect has served to inspire advancements in other fields, including medicine and robotics.
In this article, I will discuss Kinect’s evolution, outline some of the limitations that have been overcome, and provide a few examples of how Kinect is spawning advancements outside of the gaming arena.
Kinect is an impressive package of both hardware and software, which complement each other beautifully. The hardware consists of a sensor bar that comprises 3D depth sensors, an RGB camera, a multi-array microphone, and a motorized pivot, as shown in Figure 1. While the hardware has not changed since Kinect’s launch, its software receives frequent updates.
Figure 1 The Kinect sensor
Kinect’s software is what enables Kinect to track up to six players, including two active players, in real time without anyone wearing sensors. This is an extremely challenging feat that had been tackled by researchers for more than 20 years before it was cracked by Kinect developers.
Kinect’s revolutionary software is built on the concept of machine learning, a form of artificial intelligence. Once Kinect sees you with its camera and sensors, it matches up its image of you with images stored in its database, which it then uses to virtually construct your skeleton and joints to enable it to track your movements in real time, like the example shown in Figure 2. This method ensures that people of all shapes and sizes can be tracked with the utmost precision.
Figure 2 As shown by the purple lines, the Kinect sensor creates a virtual skeleton to keep track of your movements
Another key feature of Kinect is its capability to recognize and respond to voice commands. Although voice-recognition systems abound, they can be temperamental when ambient noise levels are high. Kinect’s multi-array microphone enables the Xbox 360 to conduct acoustic source localization and ambient noise suppression, permitting Kinect to recognize a voice even when the environment is noisy or the speaker moves about. This feature also makes headset-free party chat over Xbox Live possible.
Besides voice recognition, Kinect incorporates face-recognition technology. It does this by using the RGB camera, which requires sufficient light to function properly; and the 3D depth sensors, which use infrared light to measure depth, position, and motion. Kinect’s face-recognition technology enables Kinect to automatically log you into your Xbox Live account; all you need to do is stand in front of the sensor.
There was some controversy when Kinect was first released that it had trouble with recognizing darker skin tones, but this seems to have been more of an issue with inadequate lighting than flaws in its algorithms.