We haven't released a version for ARM devices yet. And when we do, it seems unlikely that we'll be able to support devices as underpowered as this.

Ok. thanks for the reply.
My Aim was to make a big Television, touchscreen/gesture recognizable, so that it can be used to play games (custom developed by our company) using gestures.
The games were very low graphics puzzle games.
Can u suggest a piece of hardware which can connect to leapmotion, run windows/linux/android.
PC/Laptops are too big. I need something small.

thanks for the reply.

What about the Intel NUC (and comparable devices)?

Those should work.

Note that the RaspberryPi foundation claims that its:

Overall real world performance is something like a 300MHz Pentium 2

so it wouldn't be too hard to improve on that (unless you were also trying to keep the price to $35).

Thanks Oliver.
I didnt knew about this. Bit costly though. :)

Hi,

I wasn't able to find a price for the NUC with a quick search of the link provided but the comments after that, paired with my knowledge of Intel, tell me it's probably way out of the price range of the Raspberry Pi. If price happens to be on of the primary reasons you chose the Pi as a development platform, you may want to consider the recently released Beaglebone Black. It's another ARM based system (like the Pi) but much more powerful than the Pi and only around $10 more. I have know way of knowing if even that would be powerful enough, but it would have a better chance of running it at around the same price-range.

Right now we don't spend a lot of effort on PCs/Macs outside of our recommended system specifications.
We are working on ARM support, however, this is intended for high end mobile applications; rather than low end embedded platforms.

In short; there is not currently any options for embedded development, sorry.

Hi mmckenzie

Leap is a startup and resources are limited, so I can understand their reticence in supporting many diverse platforms.

And, personally, I would like to see those resources applied to improvements to the current product.

I looked into Oliver's suggestion for an Intel NUC.

The price for the i.mx6q is £130.00 or over US $200. This is for an eval unit with 1.0 GGz CPU

http://shop.directinsight.co.uk/catalog/tritontxek-evalkit-for-imx6-incl-tritontx6q-p-2939.html

The price for an Intel NUC is US $268 and includes a Core i3 - 1.80 GHz CPU

http://www.amazon.com/Intel-Computing-Gigabit-i3-3217U-DC3217IYE/dp/B0093LINVK

So there's not a lot of difference. I think Oliver is on to something good...

Theo

You have good points Theo and Oliver - but I disagree with the statement that the imx6q is a low-end board. I didn't mention it is a quad-core cortex a9.. which means that it actually has 4 independent cores which each operate at 1.0 ghz (or overclocked to 1.2); as opposed to the NUC which has two cores with 2 threads each - so unless you do some sort of parallelism good luck actually getting 1.8ghz performance. i3-3217U Processor. Also the power requirements for the intel board are pretty high; I'm in the automotive sector - we have 12V power available and not that much of it.

Furthermore, I don't know where you were buying the imx6 - that website you linked to didn't work. This is one of the projects using the imx6 and would certainly allow people to do some really interesting things, without having to resort to using several computers. udoo. The thought of an arduino-controlled robot which can interact with the environment around it through a leap should make any nerd out there happy (including me!).

The other part of this is that modern cell phones/tablets use processors similar to the imx6, so if some effort were put into ARM it is quite likely that these devices would also be able to use the leap. Imx6 was just an example, although it actually is used in many of the cheap tablets coming out of asia and is capable of running android or linux.

I.mx6 has actually been used for vision before; it is capable of analyzing video at the same time as streaming video as shown here: imx6 computer vision, and there are comparision for ARM and x86 out there - arm tends to be slower (although not some of the new boards) but not in a completely different range; and neon optimizations help significantly when doing computationally expensive operations such as would be needed for leap.

I actually do own 2 leapmotions; I used them for a couple of days and then became frustrated with the many limitations that the sdk places on development with the aim of convenience, and the fairly limited scope of most of the games/applications in the store (although I'll admit this is improving rapidly - thanks to the developer community). It does what it does very well, thanks to the proprietary algorithms Leap has developed... but not everything. What gets me is that a startup company which not only relies completely on its developer community but has the stance that it is completely there to support them, still releases a product which is actually more closed-source than Microsoft's offering (yes, I know, it wasn't originally open). If they allowed developers to actually use the product to its full potential by releasing a small core (it could still be closed source) that allowed output of images which could be used in openCV or something similar, there would be a lot of people working on neat things in the linux community. And this wouldn't put their 'proprietary algorithms' at risk at all, which is one of the only legitimate reasons I've seen for them not open-sourcing more stuff. Just my humble opinion though and I'm sure there are many people who disagree with me.

Morgan McKenzie