(untitled)

Pic unrelated. In the planning stages of building a smart mirror, gonna use a pi 4 and start off with magic mirror 2, is there anything neat to add on anybody would reccomend?

If you have it where you spend most of your time, it would be pretty useful to have it connect to any security cameras you might have hooked up. Also, quality and compatibility can vary allot, but setting up a RGB matrix with a set of useful information you need (Clock, alarm, weather) can save you screen space. Getting an IR sensor + remote can be useful if you dont want to touch it / speak to it all the time.

A UV camera. It would make it so you can see UV damage under your skin that isn't visible in normal light and it also can help you apply things like sunscreen more evenly distributed. Also I assume it would make plaque on your teeth stand out more. Though the mechanism may be too thick I may also suggest a FLIR camera as a way to check for illness and disease such as fever and poor circulation as indicators and maybe have it automatically inform you if something seems out of the ordinary. Poor circulation in one part of the body for example may show inflammation which could be caused by an injury or infection. You could also set up motion tracking for gesture control so you don't have to touch the mirror getting your greasy fingerprints all over it or be that weirdo talking to their mirror more than necessary. You could set up something opposite side of the room assuming this is a bathroom or other small room with a wireless camera in it so you can see the back of your head in the mirror which helps fixing your hair. Also could use different shades of edge lighting to switch between to see how you look in different light conditions if you're planning on going out.

>>4167 I actually have a very close friend who is chronically ill so the use of a FLIR camera to pick up on changes in the body strikes me as very useful, I hadn't considered that at first, and I'll look into types of controls as well. >>4166 Yeah I definitely wanted to be able to save space on the wiggets so that's not a bad idea at all

>>4168 You may have to do a bit of research into how thermal tech is used for identifying problems since it's a relatively new use I think. I may be wrong but I think asymetry is part of a sign of a problem with an infection or maybe even a tumour while symmetrical cold spots is more of an overall problem in circulation. So not only could you maybe have a record of recorded temperatures but also maybe find a way to identify asymmetry beyond a certain threshold but that could be quite challenging to be that advanced. It depends on your skill level. You also could consider a way of measuring heart rate and detect irregular beats from a distance. There seems to be multiple ways this is possible but good luck figuring out how to apply it. I may be wrong but in theory you could maybe use the FLIR cam to do this I think but it would take the right programming to work. https://www.livescience.com/6693-sensor-monitors-heartbeat-distance.html https://www.hindawi.com/journals/jhe/2018/4832605/ https://futurism.com/the-byte/laser-identify-distance-heartbeat

>>4167 I didn't know you could spot UV damage with a UV camera! I knew you could check the coverage of sunscreen, but I'm interested in this. Also I've got a waterproof mil-spec keyboard sitting around which might be an ideal keyboard for a smart mirror. I dunno. I hate voice control and it feels like it could be more flexible than gesture control. Guess I'm just old-fashioned. It also has a waterproof trackpoint for cursor control.

>>4171 The MLX90640 looks like a good module. Far cheaper than most FLIR bolometer modules, has decent enough resolution. Anyone have any experience with that one?

OP here. Now, if I were to use it for recording, tracking, and medical use, in that case, I would just give it to my friend once I'm done but for the added context she has autoimmune disorders so would there be any use with FLIR cameras in keeping track of and aiding in issues like that?

>>4172 Yes UV cameras show damage as spots which can give an idea of risk of melanoma. UV cams also are used in criminal forensics to detect trace minerals on the skin and hidden bruising. >>4174 Would have to look into it since it likely depends on the disorder and I'm not a med student though I do sometimes research medicine. I do know if it's the kind of autoimmune disorder that causes inflamation that may show up as cold spots. Inflamation is an immune responce but only one of several kinds of immune responces. Different immune disorders may effect different responses. Since you are considering this I should mention the potential security issue of having internet connectivity. Even if it's unlikely to be hacked due to obscurity and if you had measures against hacking depending on your our friend's personality and yours she may wonder if you set up your own system inside it to spy on her.

>>4177 Fascinating, thank you for clarifying. Definitely wanna experiment with this now. Do you have a UV camera module to recommend? Or should I just find whatever I can get ahold of?

>>4178 Sorry no recomendations since I've not had spare cash to try any myself. If you want a cheaper entry point to play around with then if you have an old DSLR you don't use or willing to risk breaking if you mess up you can convert it to a full spectrum camera by removing the internal hot mirror filter. Take note the filter removed also will remove antialiasing and dust protection. From there you can either leave it with no filter as a full spectrum cam and put it together then use an external filter on the lense to block out IR or if you desire any of the visible spectrum as well or alternatively you can replace the hot mirror filter with a filter that better lets UV light through but not IR and anything else undesired and by that criteria it wont be full spectrum but dedicated to UV (or near UV. I am unsure). The standard hot mirror filters are mainly to block IR light since digital camera sensors on normal cameras aren't so sensitive to UV light. I am not sure it will make it so you can have the functionality of a UV camera or only near UV because I haven't personally done this either and all instructions I find only focus on (near) IR photography with a full spectrum cam. I wish I understood it better.

>>4181 It would be near-UV and near-IR, yep. Glass itself is not very friendly for IR and UV, so you should also find a lens with as thin lenses as possible for decent UV. For IR there are other caveats, your lens can have a hot spot in the center of the frame if the scene is bright enough. Full spectrum would be the incorrect mode of operation though, you'll have a pinkish mess instead of color with IR/UV mixed into each channel in unknown proportions, so using IR/UV/visible filters is a must if you want correct results.

>>4177 We talked about it and it does seem like UV for the inflammation would be a good way to go in terms of picking it up ahead of time. I'm not totally sure how I would go about doing that but I will get it figured out.

>>4182 Do you mean quartz lenses? Those are quite expensive Ive heard. And yes that is what I suggested that a lense filter is needed. As far as I know full spectrum images tend to be blurry looking because it's difficult to focus. Also the visible light makes it so you can't properly see the wavelengths you may want to focus on. If I understand correctly FLIR works different from an IR camera. It uses a different type of sensor that is more accurate for thermal imaging right? >>4187 It's IR that shows inflammation indirectly as cold spots. UV shows hematomas. Hematomas can cause inflamation but are not a sign of inflamation itself and only a few autoimmune disorders can cause hematomas based on a quick search. Here is a paper on IR thermography and inflamation. https://www.sciencedaily.com/releases/2015/10/151001095303.htm

doing a test run on my laptop as I haven't picked up the pi yet but so far I've got magic mirror 2 up and running, built a smart voice assistant for it , and have been looking into use for the IR camera, so far the project is coming along well.

>>4194 Have you thought about where your friend may use the smart mirror? Because it would be a potential issue if it's going to wind up in a full bathroom where sink with mirror above it is same space as a shower which means having to protect from humidity damage and have anti fogging which is extra work that isn't necisarry if it goes in another location away from steam. Also my bad I screwed up when I said cold spots show inflamation. They should appear as hot spots. Apparently the medical use of IR visuals is called thermography. Here is a short demo video some company made explaining the application of thermography. It starts with breast cancer but starts talking about other things after. https://youtube.com/watch?v=hFXlzbn9d9o

>>4195 It's most likely going to end up in the living room, there shouldn't be any issues for humidity so that's fine. as for the IR it's all good, glad o know it shows up as hot spots, it was an interesting watch. That said on the note of cameras, and since that's an area I'm not too versed in, what's a good way to make sure there aren't any back doors for it to be hacked, I don't want anybody being able to creep on my friend

>>4196 Well easiest is just no internet connectivity. You could still get current weather and forecast using a home weather station and gather the current time automatically using an atomic clock radio broadcast however this would mean no ability to play videos from YouTube on the display which is what many smart mirrors do since some people like watching tutorial videos that way. I'm not sure if that was one feature you considered adding. There also is RSS feeds which may be useful to some people but you would know what your friend needs more. Anyway another way for security is just have seporate circuits to run it which would make the camera unhackable being seporate not itself connected to the internet but that would ad extra cost and may or may not make operation more difficult to work seamlessly together. I assume you will put a sensor for motion or something else to detect someone on it to put the display and camera to sleep when no one is near it. Anything else for security would be on the software level I would imagine. Take my word with a grain of salt though as I lack the necisarry engineering and programming experience.

>>4189 >If I understand correctly FLIR works different from an IR camera. It uses a different type of sensor that is more accurate for thermal imaging right? ir cameras just see in the ir wavelength, FLIR uses https://en.wikipedia.org/wiki/Bolometer to "see" heat and you can buy standalone sensors but they are a bit pricey (IIRC there are used to be only 2 companies that make them for thermal imagers and they are export controlled)

>>4198 My bad using wrong terms. Near infared cameras are short wavelength IR cams. Thermal imaging such as FLIR is medium to long wavelength IR cams. Correct?

>>4197 Yeah not having it connected to the internet and taking the extra work to get it working together seems to be the way to go. friend will definitely need news feed, calendar, video, all of which it can do anyway. I got a Mycroft voice assistant build up and running for it, next comes getting it to work from the mirror.

I spoke with some other friends, and we discussed after this build making it into an actual product with with smart watches and blue tooth thermometers, log keeping, recording, and being a sort of all in one home care product and using this build as a prototype. what do you all think? and on that note would anybody in interested in working on the build with me?

>>4201 Do you already have idea of the mirror shape and monitor size in it? Coolest would be if the thermal cam when displayed visually could cover over the actual size and positioning of a reflection in the mirror from the front (option of flipping image or not) and be a full body mirror but that would be quite difficult to align and many thermal cams in an affordable range have very low resolution so you may or may not be able to have a resolution large enough to look right on a big display. This is just aesthetics though a small display for the thermograph and another for the widgets would be fine. Up to you. Apparently SparkFun has a FLIR setup for around $200 that can be used with pi but I am not sure the resulition is anywhere near enough. You may need more research on how much resolution may be required to be of any use. >>4202 I don't have the skills to help. Well with smart watch connectivity you could obtain information on pulse rate but I'm not sure if it can detect irregular heatbeats. Not sure what the purpose of tying it into a smart thermometer since a thermal cam if accurate enough can get temperature via certain locations on the body such as forehead temp. That only serves a purpose if the thermal cam isn't accurate enough or for use on someone so sick they can't get up to the mirror. Don't forget you need some way for it to handle multiple users too. That could either be through facial recognition or a lazy worksround is through bluetooth range readings.

>>4203 I'm not sure as to the dimensions of it yet although I was aiming for mid to full body mirror. As you said before if the terminal can pick it up there might not be a need for it, although for further use I can see how the thermometer could be a good idea. I checked out SparkFun so thank you for that I will be bookmarking that for later. going to crack the smart watch once it's gotten so I can build a work around in xcode (assuming I go with my friend's recommendation of going with an apple watch) to get it to link. it's it's not possible or too heavy on the hardware to compiler them all together on one pi, I was most likely going to get a couple and link them together so that each one handles it's own program.

>>4204 Actually after looking a bit more the sparkfun dev kit for a FLIR meant to be used with ardiuno or pi has a pretty bad resolution it seems after I've been looking around. You may be able to find better but it seems hard to find consumer available thermal camera cores. I suppose if you can't find a good thermal camera core available to ordinary consumers you could consider using one of those thermal cams meant to attach to cell phones. Though maybe you can find a way to get ahold of a thermal cam core if you act as if you represent a company making a prototype or something like that. I'm not sure how these things go. They don't even list prices on manufacture websites.

>>4206 Yeah the resolution isn't great but depending on what I can find it might have to do all things considered. I wouldn't really feel comfortable acting as a rep for a company, but I can try my best to find better ones, but we'll see.

>>4209 The mobile phone ones are similar price and better resolution. Problem is they dont have source code obtainable like buying a core can be offered and some of their apps can be privacy invasive. I'm sure there must be some thermal cam cores available to the public. It's just a matter of finding them. Though maybe you dont actually have to be a business but i don't understand why prices arent listed. Maybe they give discounts to people buying in mass which is why you have to request prices.

i found a proper thermal camera that would work well for the project and it's fairly inexpensive. got the pi4 today and am in the process of getting it set up so that should be going well too. will continute postnig updates as it comes along.

>>4223 Where did you find the cam and what resolution is it?

>>4224 I actually haven't seen it yet but I'm getting it from a friend of a friend who is a medical professional. that said as for the newest update I've gotten the pi4 and am working on the heart monitor part of the project, writing up the module but having a little bit of trouble as I'm not too use to coding in Json, it's simple enough but figuring out a way to get it to sync with an app is a bit difficult, I applied for the samsung health development and am trying to work that into it but there's a lot of moving parts here, might need additional help if anybody know show to get that set up

>>4235 Can't really help with the coding. I still am one of the few replying. Seems the Apple Watch may have been a poor choice if you went with it first thing. They still lack the ability to detect blood oxygen levels. It also doesn't have a sleep monitor built in it's health app but you may be able to make your own. Next one to come out should have these features though since it's being worked on i think. There are other fitness trackers and smart watches with these features. According to doctor on Quora who is a CEO of a fitness tracker only thing that is useful for an average person is blood oxygen levels but of course I would think detecting an abnormal heart rhythm would be useful as well. https://www.quora.com/Is-there-a-medical-watch-that-can-measure-body-temperature-blood-pressure-heart-pulse-rate-breath-rate-and-percentage-of-O2-in-the-blood

I am about doing somewhat something similar, but rather like a display for idk currency and bitcoin rates than insert cameras even in more devices, what would you recommend?

>>4290 >Bitcoin > what would you recommend? Rethinking your life. You can use some kind of number display for values to save power when not viewing a graph. Also a news feed of some kind since current events influence the economy either directly or indirectly.

>>4164 An IR touch screen maybe?