OntheEdgeofthemirror said:
billsalias said:
OntheEdgeofthemirror said:
billsalias said:
OntheEdgeofthemirror said:
|
|
Again the eye doesn't need to provide information for the software to guess depth, so yes you are wrong, well kinect can send more information, in terms of the actual application of the device it can do very little that the pseye can't and it doesn't hurt that kinect is good it hurts because it's not and you people fall for MSes advertisement champaign
|
I have a computer science degree and have taken classes specifically covering computer vision, I know how these things work. For that matter even basic trigonometry is enough to prove that a 2d image source with no additional data about the objects in the view does not provide enough data to compute any 3d positioning. There is no mathematical way to compute a Z value for any X,Y pair, it is impossible. What you seem to be hanging on is calling the ability to guess if an appendage crosses in front of the body depth information, and you are right under the right circumstances it can reasonably guess if an arm is crossing in front of the body. This is not what I described though, I was talking about a dance move where the distance your foot or arm was in front of your body and the angle it was at are known, and that cannot be done without that Z value. For an introduction to this concept check out triangulation.
As far as your assumptions, my opinions have nothing to do with any advertising campaign and to say so without knowing me is rude and arrogant. In addition to objective scientific facts my opinions come from hours of first hand personal experience with many games as well as second hand observations of the 20 or so people that have been to my house to play Kinect. No matter how much you want to deny it, Kinect works as advertised and many people find it fun.
And as I have mentioned before, this takes nothing away from Move, which also works as advertised and provides a new and fun way for people to enjoy the PS3.
So again I ask, why do you find it so hard, even in the face of huge amounts of objective and subjective evidence, to admit that Kinect is good?
|
All that and can't think of the simplest solutions, motion tracking, if an object suddenly becomes bigger it's moving forward if it suddenly gets smaller it's moving backwards, though with a 2d camera and no extra data it is fairly inaccurate but doable nonetheless
|
So you are suggesting that the you can compute the 3d position of the knee and foot relative to the hips based on relative changes in size of different body parts? You do realize that one reason move has a globe on it is because doing edge detection on arbitrary shapes from a 2d image is extremely processor intensive and inaccurate, right? What do you use as a baseline for the "neutral" position? Oh you can calibrate it at the start of the game, right. Oh wait, what happens if they move off to one side or they turn their body so the camera has a different angle on the player? Oh thats right the calibration is no longer valid and the tracking becomes exponentially more difficult, that is another reason move uses a sphere it has the same profile no matter what angle you view it from.
So no I did not miss the simplest solution, I actually thought through it and realized it was not a viable solution so did not waste time discussing it.
Also, addressing more of the original point, the reason Kinect is only now viable (as compared to when Sony researched it) is that the depth camera has only recently become cost effective at the consumer level and is needed to make it work well. See this article and this article that refers to it for some discussion.
So let's try again. You now have even more evidence piled on top of all the evidence you had before, can you finally just say that there is no logical reason for your refusal to admit Kinect is an improvement over PSEye and that it works well and people like it?
If not I am eagerly awaiting your next attempt at justifying your position.
