Take the 2-minute tour ×
User Experience Stack Exchange is a question and answer site for user experience researchers and experts. It's 100% free, no registration required.

Playing with a Kinect, I wonder how the depth channel could be integrated /combined into the video stream? Update: I want to present both data channels to the users. How could I design this?

What I got so far:

1. Side by side.

Not really integrated into the video stream is it?

2. Offset

Very much like Oliver Kreylos demonstrated. I don't know when this actually looks good. Do you have reasonable parameters for this configuaration?

3. Color coding

I could think of a relief map. Using the brightness from the video stream and color code the distance. What would reasonable parameters be for this be like?

4. ...

share|improve this question
1  
This is not the correct Stack Exchange site for such a question. You'll want to visit Stack Overflow (the overall development site) or one of the more specific development sites (though I'm not sure there is one more specific for your question). –  Evil Closet Monkey Jun 11 '12 at 17:45
3  
@EvilClosetMonkey I think the OP might be looking for design suggestions on how to integrate disparate data sets into a visual display. UX might be the appropriate place for that. However, if that is what the OP is looking for, the question isn't clear and should be reworded. –  mawcsco Jun 11 '12 at 19:24
    
Have you considered providing a slider that dynamically combines the two feeds (the live video feed and a live depth map as a greyscale overlay set multiply-blended), so you can dynamically scroll (with the scrollwheel) from completely video through blended feeds to completely depth map. It'd also be nice to have the ability to click and drag the feed and pivot it in three dimensions to see the depth rendered dynamically, but that may be overkill. –  Kit Grose Jun 12 '12 at 6:15
    
This question deals with visualization (a field in computer science that deals with how information can be presented), with image processing (a field in EE/CS) and with computer vision (a field in CS). I don't think the stackoverflow.com is the correct sight for this, however, dsp.stackexchange.com or cs.stackexchange.com/faq may be the correct place. (Probably DSP = Digital Signal Processing. FAQ: "Signal Processing - Stack Exchange is for practitioners of the art and science of signal, image and video processing.") –  Danny Varod Jun 14 '12 at 22:59
2  
It's not really a good fit for any of the stackexchange sites because there is no one single answer to your question, you're just asking for a list of possible uses. That's too broad for any SE site. –  JonW Jun 15 '12 at 6:39
show 2 more comments

2 Answers 2

up vote 1 down vote accepted

Option three can be used in two different ways. You can have a constant color overlay with variable transparency, or a constant transparency overlay using variable color.

Variable transparency would work well at separating foreground items from background while retaining the color of the original, but would give extremely low depth resolution. By this I mean that your 'background' distance would be fully opaque, and get more transparent as the item gets closer.

The problem is that we can only really detect, with any kind of accuracy, between 3 and 6 levels of transparency. This means that you will get general depth information, but the human eye cannot discern between subtle levels of transparency the same way it can discern subtle shades of gray. So the perceptual resolution is far lower. That makes this method excellent for integrating the data into one view, but poor if subtle differences need to be seen.

A constant transparency with varying colors, as in a relief map overlay, would often allow users to perceive depth more accurately (all the red items are at the same depth, etc), but would obscure the original video more. It would be harder to tell what the original colors looked like, obscured as they were by a veritable rainbow. I have not tested it, but I would not be surprised if it took more training to translate colors into depth; this may work better for a professional, but not as good for a beginner tool or demonstration.

So it depends on what you need. Both visualizations would reasonably well depending on what is important to you.

share|improve this answer
add comment

Some useful ways to combine the depth channel into the video (or information acquired using it) would be to:

  1. Segment objects so an overlay of information regarding them can be used.

  2. Segment objects so they can be kept and have the background replaced.

  3. Convert parts of the video into 3D (parts that aren't hidden from camera) and use a 3D display.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.