Sequence 01_2.mp4

In this sequence of videos you have the following sectionss, some of which may be more useful than others

The first sections shows the kinect tracking the skeletal movements of two people (you can see myself and my partner in the bottom right yellow window, messing around), its a little buggy, this can be very much alleviated by using a slightly better 3d model and some extra coding to prevent joints in the bone structure getting misaligned. This is however a limitation of the kinect in this case, as you can see, if the limbs go off frame or behind an object they will behave strangely. It can be mitigated and this could be useful in some cases, particular in things using the upper torso

The second section shows a bone structure directly correlating with real world movements, including change in the position of the center of mass, again if limbs to not go out of frame and are not obscured by another object, this is very useful implementation - the skeleton structure could be inside all sorts of 3D objects including human models. It does become clear that the kinect eeds quite a lot of space for these kinds oif things

In the third example a ball can be 'activated' by a slight push at the screen by your hand, then the ball will follow your hand directly through 3D space

In the fourth example a 3D model is being created by the depth information supplied by the kinect, its a bit rough in this case but the model can refined by smoothing, removing unwanted edges and so on - like most webcams the webcam built into the kinect (which software is matching here with its depth sensor) creates poor quality images in low light, however the depth sensor is not affected by darkness (Althugh very bright light can interfere)

In the 5th example an 'avatar' style model reflects user movement amongast many skeletons, and this could be used to modify certain movement while retaining the original gist. Because the connect can track two or more people, each person can have a skeleton, within reason (due to physical space limits)

In the final example is the kinect simply following the center of mass of people in the room, so it can track present people like a radar, this top down perspective capures my partner and i being dorks in front of the connect

Category:

People & Blogs

Tags:

• Sequence
• 01
• 2