Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: FPGA Ultrasound Imaging on a Raspberry Pi

un0rick.cc

Post reply on HN
Print view View original

31–36 of 36 posts

Re: Show HN: FPGA Ultrasound Imaging on a Raspberry Pi

#31

Post by Gibbon1 »

candiodari wrote: post #14
temac wrote: post #9

Earlier quoted context omitted.

Today a graphic chip is also good to compute ultrasound images. I suspect that you can even do it on a CPU. I'm not sure if it is that practical to process in a FPGA if you want to compute non-potato image quality: you would need to handle tons of data by the FPGA, so either use a gigantic one (with a ridiculous price) or somehow page the data (but that's quite complex to do then...) -- this seems way more difficult…

Not really: people have reacted to more capacity being available by getting in more channels. You can control a 1980s 20 channel ultrasound using even an atom cpu, not a problem (though will still get close to max cpu capacity). So what did we do ? Well we made chips, I believe the best one now has 4000 channels and there's test models of 8000 channels. A quad xeon can't process that. But it gets worse. Because infor…

" For instance, you want to create stuff to interface with ancient equipment. Lots of that is necessary, and lots of it is custom. For telecom companies, banks, factories, ... here CPUs and the fact that you can do one at a time are tough to beat. Sure FPGAs can do it better, but that's not the point here."

Guy I worked with had a recurring gig interfacing FPGA's to older Telcom equipment. Multiplexing multiple gigabit Ethernet channels into one very high speed channel which then was fed into a microwave up converter. Last and final one they did the FPGA put out like 25W. I can't imagine doing that with a CPU of any type.

Re: Show HN: FPGA Ultrasound Imaging on a Raspberry Pi

#32

Post by agumonkey »

cpfohl wrote: post #13

Seems worth mentioning that incorrectly used an ultrasound can burn bones. (Specific frequencies in bony areas moved too slowly). I didn't do enough research to understand this (but just enough to verify it). I learned it initially from my wife who's an NP. When I mentioned a previous DIY ultrasound article to her she looked skeptical and mentioned the risk of bone burns.

funny, while trying to learn more about that, I find this https://www.verywell.com/ultrasound-for-faster-bone-healing-...

:)

Re: Show HN: FPGA Ultrasound Imaging on a Raspberry Pi

#33

Post by wcunning »

zxcmx wrote: post #19
wcunning wrote: post #8

Earlier quoted context omitted.

As in all engineering related choices, there are tradeoffs that make the decision between FPGAs and DSPs: Similarities: ->Both are specialty items, with specialty dev. environments and proprietary libraries. ->Both are designed to do fast processing of highly parallel data (DSPs are known as VLIW processors, at least in all of the architectures I'm familiar with). ->Both require knowledgeable specialists to really ge…

One thing that might tip the balance in 2018: there's a growing familiarity and body of open source knowledge now around fpga signal processing among the sdr community. There are now a lot of well documented open source fpga-based data acquisition boards being sold today at comparatively rock-bottom prices - we just call them sdrs. The bandwidth requirements for ultrasound would seem to fall easily within the open so…

I'd be curious if you could point me to some links? Feel free to dm me, I'll stick my email in my profile.

Re: Show HN: FPGA Ultrasound Imaging on a Raspberry Pi

#34

Post by cpfohl »

carbocation wrote: post #16
cpfohl wrote: post #13

Seems worth mentioning that incorrectly used an ultrasound can burn bones. (Specific frequencies in bony areas moved too slowly). I didn't do enough research to understand this (but just enough to verify it). I learned it initially from my wife who's an NP. When I mentioned a previous DIY ultrasound article to her she looked skeptical and mentioned the risk of bone burns.

I have never heard of this being an issue for diagnostic ultrasound. Since I’m trained in (cardiac) ultrasonography, I think this would have come up... I also did a brief lit search and didn’t identify any obvious examples. I suspect your wife is referring to therapeutic ultrasound, which can damage bone [1]. But this is not the ultrasound people are getting of their heart, abdomen, or pelvis in general. 1 = https://…

See! I learned something new today! :D

Re: Show HN: FPGA Ultrasound Imaging on a Raspberry Pi

#35

Post by daxfohl »

analognoise wrote: post #20
stefan_ wrote: post #12

Earlier quoted context omitted.

You don't understand. FPGAs are not used for their computing prowess, but for the ability to do highly timing sensitive parallel data acquisition. At 64 Msps you have Their most important feature is determinism. You can reason about and get an upper bound on how long something takes even while programming them.

Determinism is certainly a huge part of it, but FPGAs are insanely badass at computing prowess - you can create a custom architecture to execute whatever it is you're trying to compute.

Well kinda. The flexibility comes with a cost. Clock speeds are lower and for the same logic it's going to be less power efficient than a GPU or CPU or ASIC. Logic element density will also be lower so it'd be impossible to reproduce the entirety of a GPU or modern CPU (huge caches etc) on a similar-sized FPGA. I'd also imagine that the physical distance of gate paths are highly optimized in a CPU such that more gates can be traversed per clock tick there than in an FPGA.

If you have very specific compute pipelines with "colocated" (there's probably a better word -- basically data directly on the wire, not in memory lookups) data in parallel an FPGA can offer very high compute throughput.

But if you're having to do memory lookups and such for your algorithms then the memory bus is going to be your bottleneck and most of the FPGA's advantages are out the window. At that point sure you've got a few places where the FPGA does a few operations at once, but ultimately you just have an early 1990's speed processing pipeline with a few hardware-based micro-optimizations. A GPU and its custom designed baked in logic around memory buses and caches and such are going to blow it out of the water in terms of logical efficiency, power draw, heat dissipation, logic element count, and clock speed. Total throughput will be several orders of magnitude greater.

So, it depends what you're trying to do. It's dangerous to say FPGA is always badass. Sometimes they are, and sometimes they're not even close.

Re: Show HN: FPGA Ultrasound Imaging on a Raspberry Pi

#36

Post by prueba3 »

lostlogin wrote: post #22
carbocation wrote: post #16

Earlier quoted context omitted.

I have never heard of this being an issue for diagnostic ultrasound. Since I’m trained in (cardiac) ultrasonography, I think this would have come up... I also did a brief lit search and didn’t identify any obvious examples. I suspect your wife is referring to therapeutic ultrasound, which can damage bone [1]. But this is not the ultrasound people are getting of their heart, abdomen, or pelvis in general. 1 = https://…

US Bio effects of diagnostic US are discussed at (painful) length in undergraduate radiography training here in New Zealand. We are fairly well aligned with the UK in our training pathway but in my (limited) experience echo techs don’t seem to come from a radiography background while ultrasound ones do. Could be a local situation though. Broadly, the priciples seem to be ALARM and perhaps don’t do ‘souvenir’ scanning…

Interesting
Post reply on HN