Hacker News new | comments | show | ask | jobs | submit login
Taking the SCiO Food Analyzer Grocery Shopping (ieee.org)
95 points by arm 245 days ago | hide | past | web | 51 comments | favorite



A few problems with scio:

- It's reading from (mostly) the surface, not the bulk mass. Not great for heterogeneous things like pills

- It uses machine learning models on 10-datapoint IR reflectance spectra, meaning it's only useful in a trained regime. It doesn't give information about composition, but instead classifies a sample as a member of a pretty constrained population. So a mystery substance that can't be roughly identified ('vegetable', 'pill' etc) can't be scanned, etc. If nobody's built a model for the thing you're scanning and for the property you want to evaluate, then you're out of luck

- So, evaluating drugs ("we can distinguish fake from real viagra") is done by looking at the surface coating which typically has no active ingredients (and even if there were, the signal would be washed out by inactive ingredients). The model is basically trained on 10-100 scans of a presumed good viagra pill, or maybe 10-100 different good viagra pills if they felt like it

- Building models requires purchase of a $250 license in addition to the $250 hardware, which is just ridiculous. Of course they're doing the calculations on their servers, but it still seems really scammy, hostile to developers, and counterproductive to launching an ecosystem of scanning models. The useless 10-point IR "spectra" notwithstanding, I would totally buy one of these if you could use open data and open models supported by a public community.


I got a scio dev kit from kickstarter. I backed it because they claimed at launch you would be able scan the leaf of a living plant and determine if it was sufficiently watered. Unfortunately this functionality, along with many other promised applets, was never built and the dev forums are prettmuch dead. When I looked into building my own model, I found the time required is pretty huge. To build a remotely useful model, one would need to scan between 200 and 300 samples. That means 200-300 individually potted plants for every plant variety you want to measure!!!


Yeah I was really excited about working with it, limitations notwithstanding... Until learning they charged $$$ for a dev license, which seems like a slap in the face and a hallmark of scams. Now I'm just waiting for used units to hit eBay + reverse engineer the protocol.

I'm not quite interested enough yet to buy bare Hamamatsu (or neospectra/other knockoff?) spectrometer chips, but would totally buy a breakout PCB for one if somebody had a bunch made


About $200 for a hamamatsu breakout: https://groupgets.com/manufacturers/hamamatsu-photonics/prod...

They've just finished a round, but they're reasonably regular.


Also, ~$150 for a nanolambda nsp32 spectrum sensor

https://www.nanolambda.net


I also bought one during the crowdfunding campaign, and am pretty deeply disappointed at the difference between advertised and actual capabilities.


Regarding your last point - check out this youtube vid:

https://www.youtube.com/watch?v=SfiqdooNBb8

Cheap diffraction grating and a cheap camera; basically, the idea would be to somehow use an IR sensitive camera (most are - at least in the far-mid range), an IR source (maybe an unfocused IR laser diode?), and a diffraction grating.

Ideally for the grating you'd use one for IR - but I haven't found one that wasn't reflective-based (instead of transmissive) - there's probably a good reason (likely having to do with cheap materials and IR absorption - which is why lenses and mirrors for laser cutters tend to be pretty pricey).

Anyhow - thems the basics. Take that, get an image from the camera, run a fourier transform on the data, get the peaks, then pass it thru a trained CNN (?) - heck, you might be able to forgo the transform part and just use the data from the camera directly (repurpose imagenet or something).

Yeah - I think this whole thing could be made an open-source project; probably even an instructable...? At the very least, it could become an interesting science fair project for some enterprising kid...


> At the very least, it could become an interesting science fair project for some enterprising kid

As that "enterprising kid" who couldn't wait until the SCiO was completed, and who was fed up with his 3d printed visible light spectroscope, I put in a lot of research effort figuring out how to go about creating a portable, financially attainable NIR spectroscope. I started out thinking it would be fairly straightforward since I already worked on creating CCD sensor driver boards in my early highschool years (after all it just seemed like you'd need a linear CCD with no IR-cut filter, a diffraction grating, a prism with 95%+ reflectivity, and an IR emitter with a peak wavelength of ~920nm). I invested about 6 months working through the various roadblocks I encountered until I managed to get half-decent results from it when taking the spectrum of various salts. Still never managed to fully refine it since I didn't have the budget to (Heck the Eagle schematics I sent off to a PCB fabricator were so poorly-designed that in order to get it to work correctly it required me to solder a wire across two traces to prevent the sensor from blowing up). It's still somewhere in one of my drawers and is a fun novelty item (and it did win (second?) place at my junior science fair) but at least with my design, it really wasn't the portable tricorder I was hoping for...


There is another approach worth considering. The German research group Fraunhofer has come up with a good idea to enhance the spectroscopic capabilities of commodity smartphone cameras by shining light from the display onto the object being scanned. You can read about it here: http://www.iff.fraunhofer.de/en/press/2017/app-reveals-const...

The healthtech company I work for has expressed interest to CP about using the SCiO phone for a project. If that doesn't work out, we might replicate this work by Fraunhofer. We're already looking into it. And since it's not core to our business, if we do, we'll open source it too.


Hamamatsu (Japanese optical engineering company) is offering micro-spectrometer heads, which are quite affordable (at least those for VIS spectrum) and easy to work with and they are much more precise than other DIY solutions.


Are there competitors/shanzhaied versions out yet? A cursory google search of MEMS nir specs turns up a few hits.


There is even Si-Ware FTIR MEMS sensor. It is quite pricey (more then 1K USD) and resolution is not that great, but pricey will drop eventually and it might be good enough for some applications.


Not shanzhaied, but another alternative spectrum sensor for VIS/NIR(390nm - 1000nm) is nanolambda NSP32. www.nanolambda.net


I checked datasheet and resolution is something like 25nm, not that great for most common application. Those Hamamatsu micro-heads are about 14nm (which is still not good), commercial benchtop instruments around 4nm.


It is a bit more complicated than that.

There are two classes of problems in spectroscopy: qualitative (trying to identify what the sample is) and quantitative (deciding concentrations of one or more analyte in sample).

Qualitative is quite simple: each material have quite unique fingerprint and you match it against dataset.

Quantitative is very very tricky. You have to pick good region of spectra and you have to pick good features (say peak areas or peak high). For complex solutions you will get a lot of peak overlaps and then you have problem with matrix effects for most samples (non-linearities caused by combinations of various compounds and so forth).

Good news is that a lot of interesting things can be found in near infrared region, so you don't really need expensive diffraction gratings and delicate optics.


CNN will convolve already so fft should be unnecessary.


I realized that as I noted in the second half of the sentence...


I have no affiliation with SCiO other than I went to one of their meetups. But to counter your points:

- All spectroscopy is surface reading, with NIR penetrating somewhat deeper. But it's still up to you to prepare the sample.

- IIRC, there were 400 datapoints corresponding to 1 nm spaced wavelengths from 700-1100 nm. The models aren't super smart, correct, and need a lot of collective data to be useful. They were able to do more than you describe though (a demo they did was scan cheese for protein/fat composition, though to be fair those are trained models, probably with external data).

- It's up to you (and the other data providers) to prepare the sample correctly and test it the same way.

- Yeah, I don't know what they are thinking on this.

Don't get me wrong - I didn't find SCiO particularly useful and am not sure where it's going. But I wouldn't be surprised to see technological improvement either. That is, if they have a business model and further funding.


As for the last point, you are right, they are purposefully sending encrypted/encoded(?) data to your phone, which is then sent to their server. The raw spectrum will only be returned from the server if you have an even more expensive "researcher" account...


One might assume their business model involves collecting lots of data + using it to build their own in-house models for sale to corporations, or just selling it raw?


All that you say is true (I think) and yet some applications are amazing (I backed the SCiO on Kickstarter).

Namely, it can tell your body fat from a scan of your skin, and the results seem pretty accurate (same result as a Withings scale for example).


I backed it as well. I've found the body fat analyzer applet to be highly inaccurate and variable compared to other methods (resistive, calipers etc).


So, if the drug has a time-release coating (which many do) it's probably not going to work.


Interesting, though I fear if this becomes mainstream, the producers will just optimize food for good readings. Does that mean good food? I would bet not. It's the same as with roses which nowadays look beautiful, but smell like - nothing. Unless they put a drop of perfume on it (which they sometimes do).


I read about Red Delicious apples a while back. Everyone I've met hates them so why do they have such a bold name? Turns out back in the day they were delicious, but farmers have aimed towards making larger and redder apples at the expense of taste so now we have beautiful red delicious apples that taste like ass and have the texture of apple sauce.


My concern is that this will lead to increased food waste, ie. retailers rejecting perfectly edible foods that don't scan well. That's already enough of an issue with irregular shaped produce.


> While the initial applications surround food, Sharon says that the technology is not just for checking out food freshness and nutritional information; it’s good at analyzing body fat, and distinguishing real pharmaceuticals from their fake counterparts. “We’ve done a demo that distinguishes real Viagra from fake Viagra,” says Sharon. “That’s the most commonly counterfeited drug.”

Now that's interesting. I wonder if they've considered marketing it as a way of testing the purity of illegal drugs?


> The user starts out by simplifying the problem a bit by identifying the category of the item to be examined—it’s not “What fruit is this,” but, “This is an apple, is it any good?” Consumer Physics’ cloud-based software then taps into its knowledge base, for an apple, it defines “good” as “sweet” (hence the Brix measurement), and considers an apple’s typical range of sweetness based on thousands of scans. A graphic on the phone then places the apple on a quality range.

The important data and algorithms are held in their servers, not in the phone app. This gives them a lot of control and responsibility over the models, leading me to believe that they wouldn't store data on illegal drugs.

I also feel it's kind of scummy to keep this on the server. There is no reason that this functionality couldn't be offline and on the phone, except to let them charge you monthly for the hardware you're effectively leasing from them, rather than letting you use the device you own.

Also, I expect if this does take off that someone will replace their app's phone home behavior and sever-side components with an equivalent on localhost. Slightly later will be the cloned scanners that plug into the app. They should focus on providing higher-quality and higher-accuracy scanners than competition can achieve, rather than trying to lock out competitors.


I mean, somebody could just build a model in which "Cocaine" is renamed "thiotimoline" or something, unless they already have data on known illegal drugs. On the other hand, government agencies will use better equipment so the only way that would happen would be if they were specifically trying to catch people using their device on Bad things


It could also be good for cosmetics, as they're heterogeneous and probably pretty consistent. Although it would be limited to checking brand authenticity rather than interrogating composition


Mentioning illegal drugs sounds always like a shady money grab is coming up. However, that's where this could truly safe lives.


Or perhaps testing strains of mj (in the "legal" states - medical and otherwise) for potency or whatnot?


this article is worthless "I can’t verify the accuracy of what I was seeing" ... "But it certainly seemed real" when did the ieee become a clickbait, advertisement machine.


Exactly my thought as well. Hacker news isn't very discerning with this one.


While initially skeptical myself, I found a bit of information on Consumer Physics' site stating that they use near-infrared spectroscopy. Doing a little initial research, it seems plausible that they could glean a number of useful traits from agricultural products, including soluble solids (i.e. sugar content in fruits), acidity, moisture content and more.

https://www.consumerphysics.com/myscio/technology/

https://en.wikipedia.org/wiki/Near-infrared_spectroscopy#Agr...

http://www.intechopen.com/books/developments-in-near-infrare...


NIR could give information about the scanned surface, though in this case it's apparently a low quality/low signal NIR sensor. For this app it was originally pitched as a magic device that could identify anything, but in the shipped product you not only have to tell it the category of item you're scanning, in many cases it gives very little information (e.g. doesn't identify an orange as an orange), and that information seems highly suspect and of questionable accuracy regardless.

Everything about this campaign has been dubious. One of the first videos of an "unboxing" was a close friend of the creator who, after deleting his video, went around posting comments about how it was just "good fun" and "innovation takes years", etc. Every demo of the product is a staged example of very low hanging fruit, pardon the pun. The actual utility of this device seems deeply suspect.


Of course its all "plausible" and we all yearn for a Star Trek Tricoder, but spectrometer measurements have to be taken under controlled conditions with careful calibration or else you're going to get garbage-in/garbage-out.

That said, it sounds like harmless fun.

But if you really care about accuracy, its not that hard to assess the quality of fruit and vegetables with your own senses, don't need a spectrometer/iphone contraption.


The article says "Consumers in China, Sharon points out, are particularly interested in checking food safety, given the history of problems with the food supply." Is there any reason to think that the most common problems with food safety would be detectable with this sensor?

I don't doubt it's possible there could be some correlation, but it seems like a big leap from estimating carbohydrate content to detecting food-borne illness or contamination. A quick google turns up this comment[1] (of uncertain reliability!) saying most food safety issues are caused by toxic animals or plants, pathogenic microorganisms, or chemical contamination.

[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4500434/


Most nasty stuff is present in vanishingly small quantities compared to all the other crap in biomaterial, which is why most contaminant testing involves an extraction step, or even extraction+fractionation, before spectrometry.

Somewhat relatedly, inorganic ions don't have nir activity, only their complexes with organic counterions have nir activity. And there are like a gazillion different things that complex with lead/mercury/etc. in unpurified biomaterial, so direct nir measurements of a sample are not a productive approach.

That being said, it would certainly be a big moneymaker if placebo contaminant detectors became a thing with Chinese consumers like surgical masks for smog


I am deeply skeptical of this product and find it bizarre that the IEEE would have an article with so little regiment -- basically a hyped up PR piece. Why wouldn't the author use the device themselves, without someone with a profound bias controlling the show?


The only solid info I seemed to get out of this article, is that it's a sugar-content analyzer paired with a food database.


It's not a sugar-content analyzer -- though it can do that. It is a consumer grade NIR spectroscope that pairs with a smart phone. Whether that will prove to be useful or not is up for debate but the capability for consumers seems truly novel to me.


This article is damning the product with faint praise. It's basically worthless for the purpose they use it.


Awesome! I can see hobbyists buying some of these then setting up online review hubs for each city, labelling the quality of produce in each store / farmer's market produce section when they go out on their 'measurement' excursions.


except the values will change from day to day as providers change their source mixes. nobody really cares what the exact sugar or protein content of their fruits is. It doesn't correlate with "quality".


> nobody really cares what the exact sugar or protein content of their fruits is

I disagree here - as someone on a ketogenic diet, the sugar content of food is extremely important to me. Labelling standards are abysmal, too; it's legal to market something as "0g carbohydrates" if there is less than a 0.5g carbs per serving. Considering that serving sizes can be very small, that can be extremely misleading.

For instance - a bag of pork rinds may be listed as 0g carbs, even going so far as marketing that fact on the front of the bag, while the third ingredient by weight is sugar. If the bag claims that there are 20 servings, then the only real information you have is that there are less than 10g of carbs in the bag. When you're on a diet that limits you to 20g per day, that's a huge problem.


This device does not have the accuracy to provide you with the information you need.


I think most vegetables covered in some protective layer, they don't even smell. You have to cut them open to properly use spectrometer in that case you can just taste them.


There is another alternative that looks like more of a solid tool. www.nanolambda.net

anyhow spectral sensing could be one of the main stream for IoT sensing


haha, funny layout on big screens: 3/4 of page filled with other articles.


Yep, perfect use case for 'Click to Remove Element' extension.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact

Search: