Human Locator
- Started
- Last post
- 50 Responses
- ********0
huh? car?
- Freeset0
Thanks you for your interest, at least, in our product design. :-)
The products works. The design doesn't seem to... well too bad! I still like it, but I'm questionning its use with your comments.
Surprise: a real designer, many times awarded internationally, did this job. Actually, the client (me) asked him for something trashy, new, refreshing... do what you want king of thing... and I liked it a lot.
My problem is a still like this 1st draft logo.
At least we have tried to bring something new and had fun doing it.
Now I try to copy-paste every of your comments in a word page to do a review with the designer... that is so long with this interface...
- ********0
i still wanna know how you deal with multiple target in multi-variant environment.
i mean i wouldnt believe it until i see the product in physical sense.
the flash mock up doesnt really do it for me
- Freeset0
It took 2 years to come up with this technology. A ex-guy from Sony Tokyo lead the project.
We developed for an art exibition that we are presenting in Montreal in October. None of the technologies available were flexible enough to work with video projections, text animations, sounds...
Yes, it tracks many people movements based on color. With the log, you can control Flash (it could be also Director). It's easy to have, people coordonates 30 times per seconds, their speed and their direction. Art and architecture shall benefit from it a lot. Advertising could be really effective with it... but as you said... at the end, it all comes up to the creative piece itself, not the technology.
- ********0
when i did my research in multi object targeting multi varients in complex environment, the color can cause the complex readibility issue.
Also tracking multi target in complex flexible setting causes issues of in-ability to track targets correctly.btw this is not a new technology thathat has been implemented. Barneys NY had light sensor that did that same thing about 3 years ago. Which was amamzing but that system also fell in to problem of multi targeting.
i could see how you guys are using positioning quads to track varient in color and movement.
Have you considered color deept or environmental effects? such as sunlight?
if there are 20 people walking around at the same time. there is no system capable enough to track in color since, objects moving to quads will maybe very well have similar color.
if you have done chao test on the system, it it likely that it fails to track proper targets.
- Freeset0
Now these are relevant comments. Yes it wasn't easy, and it will probable take 10 more years before to come up with a perfect technology. But still, Human Locator deals pretty well with the 2 issues you mentioned.
1) Multi tracking: works fine. More important is the effect it generate on the creative piece - will it change something if there are 21 people instead of 20? Generally, after 5, it doesn't really matter. Anyway, Human Locator can be program to consider groups (if 1 to 4 people do this, if 5 to 10 people do that, if 11 to 20 do that...)
2) Working in sunlight or in environments with a lot of reflexions is more difficult. The colors change every seconds, but still, the Human Locator manage with that. It took two years to develop. For professionnal results, we can add a human supervision based on a red alert system. This helps in major situations and to better develop our product.
- ********0
umm that really doesnt tell me anything..
its just another product overview
- guardia0
Hi,
To reply to MrDinky, I'm the one that worked on the technical stuff.. At the moment, the system tracks blobs, and uses color information to track it around the image and differentiate it from other blobs when they cross. It doesn't work very well (read track, the system is aware it cannot track anymore) over 5 targets at a 320x240 resolution, but this is being worked on... because of that and lightning condition problems in changing daylight for example, we are working on another system.
- ********0
i think using higher resolution works better with fast fps camera. if you can consolidate color information, it can work better or tracking color ranges.
the lighting aspect of it is crucial. BUT can be worked around (adjusting contrast data).
one thing you might encouter in an outdoor setting is changes in background.
lets say if you put the system by the windown with car traffic.
thats gonna be a pain in butt. you have to know what is target and ignore other movement as part of back ground..best of luck with you.
hope to see the final product when its done.drop me an email if you need an "opinion" on anything.
- JamesEngage0
it actually sounds quite interesting, but your website does you no favours!
Especially when you consider it doesn't differentiate from a interactive Christmas tree I did at Olympia in London, and got a techincian from the Mill (was working there then) to rig up the sensor for when people walked past.
I think you need more research into what/how you actually trigger a response as much as what the technology is.... you results don't seem to offer up what you could do if 1-4 people/ 5-10 people are watching etc?
Perhps you need creative solutions as to what the technology can do given it's restraints or more importantly from what actually controls the system.