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.
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 I tried to get their SDK running and their own documentation was out-of-date, missing necessary links, and sometimes blatantly wrong.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.
Not discrediting the article however is this partially incorrect? Most of these embedded Linux devices that are “intelligent” doorbells, that is run inference locally vs cloud compute, can’t locally build models however can run inference. So they must send new facial images of people they want to identify to a cloud solution to do so?
Long day, not sure if words are my friend right now…
there seems to be two main issues: (1) the uploading of thumbnails regardless of user settings and (2) the ability to watch a video stream unauthorized. the discoverer only details the first, to give anker time to investigate and fix the latter. besides their potential use for image recognition, the thumbnail images are used in notifications. i don't think any devices are powerful enough to do facial recognition on device yet. we've just started to get voice recognition on device, and that entails much less information and a lot smaller problem space. but if someone says they don't want to use cloud services, no data at all should be sent to the cloud.
"With regard to eufy Security’s facial recognition technology, this is all processed and stored locally on the user's device."
Running a simple face recognition algorithm locally seems entirely possible. Just measure distance between the eyes, distance to nose, etc., and keep a local database of all the faces encountered. Most of the time the basestation is not doing much except storing the video stream, so it may have enough spare capacity (say capacity that would otherwise be used to stream video to a phone) to also train an AI in the background, even if it takes 30 minutes per face. Whether that database/training gets sent back to eufy and is added to a global face recognition algorithm is a different matter. Running an AI can require very few resources depending on the algorithm. Training takes more resources but can be run over the long term. Smartphones are perfectly capable of running and training a variety of ML algorithms and a basestation that can handle many video streams is likely to be more powerful than a smartphone. They might even outsource the computation to your phone.
Nest cameras claim to do "ML at the edge", and then there are devices like nVidia's Jetson Nano. I strongly believe that facial rec without going to the cloud is possible with current tech.
Guessing here, but I can see propagating those ML models, where one camera learns of the face and it is sent out to other cameras in the same "security system" logical unit as a reasonable thing.
IPhones do local facial recognition in the Photos app. I don’t expect a security camera to be as powerful as an iPhone but if that is the main thing it is doing, it is not an unreasonable thing to expect it to do.
This is a difficult thing to just say "I do this and here is what I did". People's body's can be very different. For example, pro cyclists burn fat and carbs differently than normal people do to years of conditioning( aka base miles). That high fat diet only works for super intense athletes. Try and let me know how your cholesterol levels end up. Haha
Food prior to very long endurance Activites is good to stop the gastrointestinal events during the event. So I do agree with the benefits there however gels unlike solid food better absorb and replenish carbohydrate stores so you can maintain higher/intense efforts for longer. Otherwise you could "bonk". But yeah... It all depends on the activity, duration, intensity , pre-training and food consption etc...
I race/ride bikes as well. So 250-300mi or 15hr each week on the bike( more if I didn't have a day job).
I got my family's high LDL and cholesterol, which normalized for the 2 years I was on keto and stayed low after.
I went from having my GP threatening to put me on Crestor to asking for blood work every month for a while to accepting it was working and probably not going to be the death of me. Apparently dietary cholesterol doesn't really affect blood level.
I'm not an athlete, but I do about 10 miles a day. Nothing special.
IMO - This article is an over simplification. Data sheets need to be book worthy documentation due to the behavior of complex semiconductors. I believe Datasheets are law abiding documentation? If something is out of spec or out of characteristics is is considered faulty which means loss of money or competitors can get in.
What I'm really interested is in what peoples opinions are regarding the navigation of content and material of these large scale Semiconductor Sites. They have 50k plus products which are all very complicated and specific to a use-case. I find there is a constant struggle to route people to the right solution especially when engineers can use different parts in new creative ways. ADI, Texas Instruments, and Maxim all fall in this boat.
>I believe Datasheets are law abiding documentation? If something is out of spec or out of characteristics is is considered faulty which means loss of money or competitors can get in.
There are plenty of errors in datasheets. Analog Devices even has this disclaimer at the bottom of their datasheets:
"Information furnished by Analog Devices is believed to be accurate and reliable. However, no
responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other
rights of third parties that may result from its use. Specifications subject to change without notice."
TI's website is still pretty decent for navigating around and finding the parts that'll solve your problem. ST seems to be making it harder and harder. The former Linear website was very easy to use. Maxim has always sucked, in my experience. NXP has an OK website but generally quite good datasheets.
If you're designing electronics for profit (not hobby), reach out to your local non-stocking distributor (Arrow, Avnet, Future, etc) and ask to talk to the local sales person and FAE. Tell these people your problem. They have contacts at a bunch of semi vendors and the FAE's main job is to know what parts do what. A good local FAE and sales person are worth their cost when you're trying to find what parts to use, even though they get much less useful once you're into the minutia of any given part or design.
Finding good and cheap ways to solve electronics design problems really comes down to experience. Not that you've done a ton of designs yourself but more that you've looked at a variety of ways that other people have solved problems and then tried to understand why they did it that way. Working at more than just one company where you get to do design reviews is a huge help with this. Ask lots of questions at design reviews, challenge the designer on why they're doing things a particular way even if you don't know how to do it better, the designer probably contemplated a handful of ways to solve the problem and will be happy to explain their decision to you. It's also helpful if you can work with a team who are trying to build low volume (<10k units per year) products and a team building high volume (>1M units per year) as there's very different tradeoffs that engineering, manufacturing, and the business are willing to make as volumes change.
Ideally true, but in practice I’ve seen certain performance claims (such as battery cell capacity degradation in a battery data sheet) aren’t accurate or depend a lot on your application. If they are reputable, I assume they are trying to estimate an honest average.
Navigating the sites are challenging when I don’t already know which part I want, for sure. I wish there were better ways to explore sites for certain keywords in a data sheet.
Digikey has one of the better search engines, not least because you get preview thumbnails of almost everything, and you can search in single unit quantities. It also has some intelligence when filtering parts by value (eg Leds by wavelength).
The problem is usually knowing the names of things, and eventually you learn what keywords to search for eg 2row 10way for a 2x5 connector. Everything is mostly standardised.
Octopart is best for finding stock, I wouldn't use it to locate a new part.
For finding new chips, like random sensors, the best way is to trawl the major IC websites (Ti, Max, etc) and see what pops up. You can still get samples from most of these places, and of quite expensive parts too.
There's a group of wealthy people in Austin proposing a Heathrow Pod-like system. I'm skeptical that it can handle the large surge of passengers from a large music venue or Circuit of the Americas.
Typically "4 people" would equate "2 Americans". Granted, big cities tend to have lighter people than rural areas, so it is less of an issue. Or we an always American size it.
Maybe I'm crazy but if it is a physical battery swap then this brings memories of history class. Horses were swapped along the silk road to extend distance in shorter range of time. Same goes for horses as do batteries I guess. They get the job well on more available energy but take for ever to regain to keep going. 2c of thought
I really doubt horses were fungible. I don't know the actual details, but I would expect when you swapped horses, you got off a company horse and got on another company horse.
That's a TWL6030 power management chip, used for charging and powering the board from the battery. There are no side photos, but sometimes there is a trick you can do with the OMAP: put it under the RAM. It can provide pins for the RAM to be soldered on.
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.