It's easy to trick by just having your hand anywhere in front of your face, even a foot away. Is this something current image classifier architecture could address with more data? It seems like it'd be hard to tell whether a hand is large and on a face, or small and far away from the face unless there is some sort of depth estimation going on (based on cues like shadows).
> I looked up your code because you don't know how to hide it and have made my own cloned app for some reason.
Nobody liked you in high school, I'm sorry you had to find out this way.
Cool project, but what got me to comment was the really stellar FAQ on the landing page.
I think it's really hard to strike capture the tone of software in written descriptions of those pieces of software. Coding projects that are programmed in an experimental and light-hearted way can end up being spoken about in somber tones that lose that sense of fun and playfulness.
I think that this opinionated writing style is really great for product documentation, particularly when you are documenting something that is not intrinsically interesting.
As an author writing like this you are leaving yourself vulnerable. It shows that you have fuzzy, opionated and possibly irrational thinking, which is very easy to criticise. This is unlike our typical dry and academic writing where we remove ourselves and personality from our prose. I am guessing that this vulnerability is what makes you feel foolish when you write this way.
I've also been told that this style is not great for non-native speakers.
Maybe train people not to touch their face and in higher risk situations (think public transit with lots of touching or bigger group events) add a mask as another layer of protection and a reminder.
Start wearing a bracelet or something else to give yourself a little reminder when you are reaching for your face. I have no idea whether bracelets themselves are germ traps, of course.
Of course the killer app here is to use the accelorometer in a fitbit or a other fitness bracelet to play the danger ring tone on your phone if your arm comes up to your face.
How about this solution: If your face itches or something, use the back of your hand to rub it? Or some other part of your body that you can be confident isn't used to touch anything potentially contaminated.
Partition your body: some parts used for some things, others for other things. Use an elbow or knee to push buttons, use the long sleeve of a jacket as a mini-glove if you need to grasp a handle. Or wear gloves that you take off once you're inside.
I see a lot of people using their jacket sleeve to open doors or to cover a cough, etc. But how often do you wash your jacket versus washing your hands? I’d rather open a door with my hands then wash my hands (or use sanitizer if a sink isn’t around) because I can wash my hands several times a day. I can’t wash my jacket more than once a day.
Consider the sleeves of your jacket to be permanently contaminated. (Though hopefully the inside of your elbow is contaminated only by your own coughing.) Repeatedly contaminating the sleeves doesn't make them worse; just try to avoid having them touch anything else important. (Theoretically you could be worsening the doorknobs, but it seems unlikely you're adding anything they don't already have.) It may not be a perfect solution, but I think it's pretty good.
Regarding washing hands, I've found that if I do that too much then my skin (starting on the back of my hands) tends to dry out and flake off. Certain kinds of moisturizing soap are less bad at this, but I don't choose what soap is available except at home, and it has led me to find other solutions like the above.
Avoid laurel sulfate based soaps. C12 is particularly well sized to make micelles that disrupt the skin. Castille soap/ stuff made from good ol' fats and lye are the most gentle, laureth/pegylated sulfates are less-bad than lauryl sulfates.
I liked the way yours worked a little better than the dtyaface version. I like that yours allowed recording of samples since I have a three monitor setup and my laptop with camera is at a side angle. Your app handled side view face touching alerts quite well.
Congrats, you've made the least annoying one so far!
A simple beep is all we need, people. Also it seems more accurate than the others few linked here it seems, the face part is more narrowly detected, the other two went off whenever my hand was in the frame at all for some reason. This one only beeps when I actually put my hand on my face. (Not to be super critical of toy projects)
and it felt like the last one was the strongest (which is what we released). The one we built is as strong as the training you give it—it shouldn't be going off if your hand is at all in the frame.
The URL of the actual app is terrifying. I was certain the site had been hijacked when it redirected to the most suspicious looking URL imaginable at the domain "drv.tw": https://lzdmsmujepoc2xlgp13srg-on.drv.tw/Don%27tTouchYaFace/... (HN shortens this, but I think the real kicker is the '.com.html' at the end)
I mean, he was able to put that website up and running (the main url), then another bizarre sketchy URL. What's up with paying something? How much is hosting, like $2 a month?!
drv.tw is a "Google Drive to web" proxy. It lets you stick an html file on google drive and have it on the internet.
I'm not sure how it pays for itself - after all, every byte served has to go through their servers, and running a service like this I can imagine a lot of illicit content gets posted too (requiring a lot of admin time).
Great idea. Also, after people become trained to change their behavior when hearing that Ralph "I'm in danger", start playing it periodically when people are on Twitter, Facebook, etc.
This is awesome. Does anyone have a tutorial they recommend for building something like this (preferably in Python)?
https://teachablemachine.withgoogle.com is awesome and it looks super easy to use - but any tutorials if you wanted to learn it the harder way via OpenCV or Darknet?
Being on the spectrum of ADHD, I'm pretty certain that it's going to be absolutely impossible for me to stop touching my and suspect that small nervous habits of this sort just aren't going to be worked out of a significant portion of the population even if I could stop myself somehow.
I'm fortunate I can work remotely. Social distancing is a thousand times easier for me than stopping nervous habits, include touching my face. I suspect such actions are on the interface between conscious control and reflex actions.
I suppose the authorities should still keep telling us to stop doing but I think planners should pretty much assume this is not a barrier that's going to stop the disease.
I'm in a similar situation. I have a very large number of physical tics (classified as "tic disorder" and not "Tourette's" because mine are all nonverbal), and on top of that I have a massive number of sensory processing issues that make me react very strongly to any sensation on my body so I'm constantly addressing them, and I've long since decided I'd rather get the coronavirus than not touch my face. The coronavirus cause me less discomfort than not touching my face, by multiple orders of magnitude.
I have addressed this issue with a simple hardware innovation: the Facespoon.
Find a clean object with a well-defined handle. Use the non-handle end of the object to touch/scratch/adjust/manipulate your face. Clean the object regularly.
Working from home, I've designated a kitchen cooking spoon on my desk for this dedicated task. Plastic appears to better than wood, as it is more-readily cleaned and does not absorb facial oils.
Do your own due diligence on this one, but a few years ago I think it was reported that, surprisingly, wooden chopping boards are better than plastic ones (which had been assumed to be better as they're less absorbent, etc.) and actually combat bacteria.
Given Covid19 is reported to - unusually for a virus - last up to a week [check a proper source!] or so on hard surfaces then maybe in this situation wood might also be better???
>"A study by the Food Research Institute in Wisconsin
(Ak et al. 1994a and b) compared wooden and plastic
boards and came to the surprising result that wood possesses substantially better hygienic characteristics than
plastic. After contaminating different cutting boards with
bacteria, significantly fewer viable bacteria could regularly
be recovered from wooden boards than from plastic
boards. These results were confirmed by Gehrig et al.
(2000) in a recent study investigating hygienic aspects of
wooden and plastic boards regarding the risk of food
contamination. Previous studies assumed that the detected
reduction in bacterial numbers on the wood surfaces is
caused by an antibacterial effect of wood based on several
physical and chemical properties of wood. The porous
structure and hygroscopic characteristics of wood could
remove the water needed by the bacteria for their vital
functions and multiplication and thus kill them
(Kampelmacher et al. 1971, Schulz 1995). In addition,
substances present in wood (e.g. polyphenoles) could be
responsible for an antibacterial effect (Willaman 1955,
Biswas et al. 1981, Laks and McKaig 1988, Field et al. 1989,
Schra¨gle and Mu¨ller 1990, Scalbert 1991, Mu¨ller et al.
1995)"
>from DOI 10.1007/s00107-002-0300-6, "Wooden boards affecting the survival of bacteria?"
I am aware of that study -- my experience with a wooden Facespoon is that it can actually absorb some facial oil, at which time I suspect the wood is more problematic.
The difference is that plastic cutting boards get cut all the time, with the knife jamming the cooties into the grooves. So they are hard to clean well. (although if you clean with bleach they are fine, and that's what you usually have to do in restaurants....wood cutting boards are discouraged by health departments)
A plastic spoon is probably pretty smooth so easy to clean.
With all the talk about not touching your face, I fear that sometimes this is overemphasized for Corona virus. It is a virus with a viral envelope, and viruses with viral envelopes are especially unstable outside of the human body. Keeping your distance to people, not staying long in rooms where droplets may floating in the air carrying virus is just as important if not more.
Also, after washing your hands with soap, viral counts should be so low that there isn't a danger of infecting yourself anymore, so the situation where you sit at your laptop isn't especially dangerous IF YOU WASH YOUR HANDS BEFORE USING IT (and possibly wiping the keyboard down with a wrung-out cloth that was soaked in soap water).
None of that is well-established. It might be right, but there hasn't been time to do the science. Certainly facial contact is a known vector for many viruses, including other coronavirus genera.
This is a crisis. We have to work things via defense in depth and hope that it is enough. So we wash our hands, isolate, wear masks where available, and we don't touch our face.
Please don't try to finesse a pandemic just to seem smart on the internet. No one knows whether you're right or not.
If you want to actually stop the spread of airborne diseases you should wear a full face respirator. The expensive models are extremely comfortable and can be worn indefinitely.
This immediately reminded me of the joke about a tourist talking with an old Australian bushman. The original rambles on a bit, but the shorter version is the tourist asks if he knows any bushman remedies for chapped lips.
“What you do, mate,” replies the bushman, “Is lift up your horses tail, and kiss him straight on the bum.”
“And that fixes your chapped lips?!” exclaims the tourist.
“No,” replied the bushman, “But it stops you licking them.”
Dr. Michael Osterholm was on Joe Rogan's podcast explaining in detail how this virus spreads [1]
- Washing hands and not touching face is inconclusive but it makes sense to keep up the hygiene. That's not the primary means of how this virus spreads.
- Virus primarily spreads via airborne means, simply by breathing (not even coughing or sneezing) is enough. It stays suspended in air (you can imagine the negligible weight of a 0.1um particle compared to the fluid mechanical forces from circulating air).
- N95 masks (contrarily to the popular myth) are the most effective means of stopping the spread. Shortage of masks for hospital workers is an orthogonal issue.
I urge everyone of you to watch this interview and to be more informed.
144 comments
[ 3.7 ms ] story [ 204 ms ] threadIt's still a brilliant idea and I love it, regardless of its limitations.
I really felt this.
To get me to help them with their math homework.
I think it's really hard to strike capture the tone of software in written descriptions of those pieces of software. Coding projects that are programmed in an experimental and light-hearted way can end up being spoken about in somber tones that lose that sense of fun and playfulness.
Especially:
> My dude, I can barely make a website, much less a mini NSA.
As an author writing like this you are leaving yourself vulnerable. It shows that you have fuzzy, opionated and possibly irrational thinking, which is very easy to criticise. This is unlike our typical dry and academic writing where we remove ourselves and personality from our prose. I am guessing that this vulnerability is what makes you feel foolish when you write this way.
I've also been told that this style is not great for non-native speakers.
How many times did you touch your face testing it?
Maybe train people not to touch their face and in higher risk situations (think public transit with lots of touching or bigger group events) add a mask as another layer of protection and a reminder.
Start wearing a bracelet or something else to give yourself a little reminder when you are reaching for your face. I have no idea whether bracelets themselves are germ traps, of course.
Regarding washing hands, I've found that if I do that too much then my skin (starting on the back of my hands) tends to dry out and flake off. Certain kinds of moisturizing soap are less bad at this, but I don't choose what soap is available except at home, and it has led me to find other solutions like the above.
Edit: this may be a potential vector for reinforcement learning
Plus, it's all front end, so it's not hidden!
> 9. I looked up your code because you don't know how to hide it and have made my own cloned app for some reason.
> - Nobody liked you in high school, I'm sorry you had to find out this way.
.
.
.
me
A simple beep is all we need, people. Also it seems more accurate than the others few linked here it seems, the face part is more narrowly detected, the other two went off whenever my hand was in the frame at all for some reason. This one only beeps when I actually put my hand on my face. (Not to be super critical of toy projects)
pre-made models
hand and face detection separately
training the algo yourself
and it felt like the last one was the strongest (which is what we released). The one we built is as strong as the training you give it—it shouldn't be going off if your hand is at all in the frame.
Same...
> Why does it open in a new tab with a different URL?
> Setting up websites is hard and I got frustrated and you're not paying anything, so...
I'm not sure how it pays for itself - after all, every byte served has to go through their servers, and running a service like this I can imagine a lot of illicit content gets posted too (requiring a lot of admin time).
Why not just use github/lab pages?
Also terrifying... I mean, it's google... after all - Pretty sure my android device knows when I'm touching my face when it's turned off.
result:
http://win2003.verylegit.link/15notice.dmg.docm
The world will be a better place.
https://teachablemachine.withgoogle.com is awesome and it looks super easy to use - but any tutorials if you wanted to learn it the harder way via OpenCV or Darknet?
I found this tutorial from fast.ai but haven't been able to check it out yet: https://course.fast.ai/videos/?lesson=1
I'm fortunate I can work remotely. Social distancing is a thousand times easier for me than stopping nervous habits, include touching my face. I suspect such actions are on the interface between conscious control and reflex actions.
I suppose the authorities should still keep telling us to stop doing but I think planners should pretty much assume this is not a barrier that's going to stop the disease.
Find a clean object with a well-defined handle. Use the non-handle end of the object to touch/scratch/adjust/manipulate your face. Clean the object regularly.
Working from home, I've designated a kitchen cooking spoon on my desk for this dedicated task. Plastic appears to better than wood, as it is more-readily cleaned and does not absorb facial oils.
https://en.wikipedia.org/wiki/Antimicrobial_properties_of_co...
Do your own due diligence on this one, but a few years ago I think it was reported that, surprisingly, wooden chopping boards are better than plastic ones (which had been assumed to be better as they're less absorbent, etc.) and actually combat bacteria.
Given Covid19 is reported to - unusually for a virus - last up to a week [check a proper source!] or so on hard surfaces then maybe in this situation wood might also be better???
>"A study by the Food Research Institute in Wisconsin (Ak et al. 1994a and b) compared wooden and plastic boards and came to the surprising result that wood possesses substantially better hygienic characteristics than plastic. After contaminating different cutting boards with bacteria, significantly fewer viable bacteria could regularly be recovered from wooden boards than from plastic boards. These results were confirmed by Gehrig et al. (2000) in a recent study investigating hygienic aspects of wooden and plastic boards regarding the risk of food contamination. Previous studies assumed that the detected reduction in bacterial numbers on the wood surfaces is caused by an antibacterial effect of wood based on several physical and chemical properties of wood. The porous structure and hygroscopic characteristics of wood could remove the water needed by the bacteria for their vital functions and multiplication and thus kill them (Kampelmacher et al. 1971, Schulz 1995). In addition, substances present in wood (e.g. polyphenoles) could be responsible for an antibacterial effect (Willaman 1955, Biswas et al. 1981, Laks and McKaig 1988, Field et al. 1989, Schra¨gle and Mu¨ller 1990, Scalbert 1991, Mu¨ller et al. 1995)"
>from DOI 10.1007/s00107-002-0300-6, "Wooden boards affecting the survival of bacteria?"
Viruses have a lipid outer layer, but no metabolism of their own. Sometimes dehydration will break the capsule, sometimes it won't.
Soap is usually quite effective due to saponification physically destroying the virus's capsule, fwiw.
A plastic spoon is probably pretty smooth so easy to clean.
Also, after washing your hands with soap, viral counts should be so low that there isn't a danger of infecting yourself anymore, so the situation where you sit at your laptop isn't especially dangerous IF YOU WASH YOUR HANDS BEFORE USING IT (and possibly wiping the keyboard down with a wrung-out cloth that was soaked in soap water).
This is a crisis. We have to work things via defense in depth and hope that it is enough. So we wash our hands, isolate, wear masks where available, and we don't touch our face.
Please don't try to finesse a pandemic just to seem smart on the internet. No one knows whether you're right or not.
Seriously, just don't touch your face.
Here's hoping they become acceptable.
“What you do, mate,” replies the bushman, “Is lift up your horses tail, and kiss him straight on the bum.”
“And that fixes your chapped lips?!” exclaims the tourist.
“No,” replied the bushman, “But it stops you licking them.”
- Washing hands and not touching face is inconclusive but it makes sense to keep up the hygiene. That's not the primary means of how this virus spreads.
- Virus primarily spreads via airborne means, simply by breathing (not even coughing or sneezing) is enough. It stays suspended in air (you can imagine the negligible weight of a 0.1um particle compared to the fluid mechanical forces from circulating air).
- N95 masks (contrarily to the popular myth) are the most effective means of stopping the spread. Shortage of masks for hospital workers is an orthogonal issue.
I urge everyone of you to watch this interview and to be more informed.
[1] https://www.youtube.com/watch?v=E3URhJx0NSw