[schreiaj]

What should we run it on if we want to deploy it on mobile systems with limited access to network and extremely restricted power budgets?

For a hobby project I'm trying to find a solution - Power budget for multiple is sub 200w. Need to run inference on a lower resolution video stream (or multiple, that be nice) to do object detection. Cost is a factor because I need to have multiple angles to determine where in relation to a mobile platform the target object is. I'm looking at the Coral.ai board because RPi like boards lack the ability to do ML tasks at reasonable FPS and NVidia seems to have abandoned the lower cost side of the market since the Jetson Nanos seem to be less and less available. (Not that Coral.ai boards are available at all...)

[Eisenstein]

Check out the Luxonis Oak products. I use an Oak-1 Lite to do real time 2 stage object detection and recognition (~23FPS at 1080P inference on device with two custom yolov5n models). With a bit of python and a Pi (or a Rock64 or similar) you can get it up and running in a day. They also have a decent community and are actively developing the API/SDK and hardware.

[schreiaj]

Thanks, I've got one of their depth cameras that's been ok. I didn't realize they'd expanded their line so much. Glad to hear about the API/SDK improving, last time I mucked with it a year or so ago it seemed like it was underdeveloped.

Going to have to dig into the sensors they use - had passable luck with non ML tasks using dirt cheap camera modules from laptops running at low resolutions right up until I started moving the cameras at all and then it became a blurry mess because they were so small their exposure times were high. (I'm trying to also avoid having to put a bunch of illumination near the cameras so it doesn't entirely look like a biblically accurate angel)

[Eisenstein]

Well, you can control the camera ISP and they use very decent IMX modules, so it really shouldn't an issue like it was with the cheapos, as long as you can do the coding to your needs.

EDIT:

* https://docs.luxonis.com/projects/api/en/latest/samples/Colo...

[eisen]

Check out TI’s new low cost board vision AI board: https://www.ti.com/tool/SK-AM62A-LP

They have a home grown AI accelerator along with free Deep learning SDK.

Also offer a pretty easy online tool (free again) to use called TI Edge AI studio. They are using extending the existing AI solutions that come from the higher performance parts like TDA4 and AM68A parts. Pretty good considering a lot of these manufactures are just buying other AI Ip that isn’t performing great and investing in their own engineering.

[Eisenstein]

Looks nice. Honestly I am just kind of playing around right now and the Luxonis products are the only ones that seem to have any kind of active development, support, and usable (for me) hardware in the hobbyist (<$200) price range.

nVidia's platform is just a huge mess. I tried to get their SDK running and their own documentation was out-of-date, missing necessary links, and sometimes blatantly wrong. I wasn't going to dump $400+ into that ecosystem.

Google gives up on hardware consistently and has the worst support of any existing software company (effectively zero) and has bungled the AI hand every change they get.

ARM NPUs I am not going to bother with. I can't even get video encoder acceleration working on a non-Pi ARM SoC except for the Rock64 and that is like 6 years old and was missing that functionality for 4 of them.

Intel only cares about its corporate partners and doesn't give a crap about hobbyists in regards to A.I. But their VPU was (is) decent and Oak guaranteed supply for at least until 2025 or thereabouts and built a useable API so we don't have to mess with OpenVINO.

It's all a mess right now but I can't say that competition is bad. It will be nice if we dispense with all the bespoke platforms and agree on some common architecture for edge devices, but I won't hold my breath.