Nude.js | Nudity detection with JavaScript and HTMLCanvas
patrick-wied.at
patrick-wied.at
You know, things people should actually be concerned about children seeing.
For one subset of violence, gore pictures, you could identify body parts (already solved) then look for what appears to be wounds using color changes and edges.
That's of course one small part of the concept of violence - you're totally correct in stating this is difficult.
Also hateful is more objective/political so I don't think we have much chance there.
I would wager it would be rather easy to make a system that detects bloody gore at a higher accuracy rate than this crap. Of course I have little interest in aiding any sort of censorship, so you won't see me attempting it.
I'm not seeing either statement as more legitimate than the other. We all know porn is being singled about because of an absurd prudish attitude the west has towards nudity.
So you don't mind people jerking off whilst you're trying to enjoy a coffee and a cake or get a book in the library with your kids?
I mind, call that absurd if you like.
It's not nudity that is the problem in public spaces like libraries it's use of pornography. A simple way to avoid most masturbatory use of such spaces is to prevent people from being able to view pictures of other people who are naked.
Even if it were prudishness, then what right do you have to dictate to the "prudish" what their attitude to nudity or sexual imagery in public spaces should be?
We have public indecency laws to cover that. There's no reason to place further restrictions.
There are laws against car theft as well, but it's much easier to lock the car and set an alarm.
And as for the mentally ill? Well if walking down the street of your average sized city is anything to go by, lack of pornographic material isn't much of an inhibitor.
Maybe because the homeless or impoverished get internet in libraries and nowhere else, but regardless of the cause, something should be done.
The last time I used a computer in a library, I reached down to grab the side of the chair to scoot it under the desk, and received a handful of ejaculate for my trouble. It was a nice library too.
One of my best friends works in a different, nicer library, and constantly complains about people jerking off in there. They catch them and kick them out, but it doesn't stop the problem any more than swatting mosquitoes stops that problem.
You may say mental illness because of your particular environment or upbringing, but I feel below a certain level of society it's the natural state of people-- if they're able to see a naked picture they pursue the pleasurable sensations it produces until that process completes itself. Altering the environment seems like a damn fine idea to me.
I looked at my hand as awareness set in, and then I stood up and walked to the bathroom. Then I told the librarian, and she gave me some paper towels to clean the chair. I couldn't think of how getting angry would benefit me so I didn't bother.
While image detection algorithms aren't deployed widely, I'm sure many public spaces and internet cafes use some type of blacklist filter for 'known' porn sites.
You're being disingenuous - all public computers where I am have some sort of filtering as to the websites you can view. Hence preventing those with low social inhibition (for whatever reason, horniness is strong driver) from seeking out porn in those places.
Side note, there really are more valid reasons to block pornography than kittens. Preventing access to pornographic materials on a group of managed machines is just good security policy, for many reasons.
2) Child pornography. Sure, there may be laws that protect you if it's discovered on a hard drive of a machine you manage, but I'd much rather avoid that mess all together.
2) If someone is crazy enough to download child pornography onto a public machine, they're almost surely good enough to bypass whatever protection mechanisms you may have in place.
2) What does crazy have to do with skilled?
That's a really good question -- the short answer is it's a really great motivator. Some random pedophile, who wants to obtain child porn, first must learn how, and in the process they would have to learn some basic computer skills, as well as learning to bypass the mechanisms we (rightly) have in place to prevent such material from being distributed, and to catch those who download it. This is why malcontent teenagers who would want to do something like DDoS tumblr are far more likely to know how to do that than well-adjusted teenagers who don't care.
And seriously, how often does that actually happen? This is just another 'think of the children' argument designed to disable critical thinking.
And as for your first (uncited) assertion, consider the following: http://www.theregister.co.uk/2010/06/30/unsafe_surfing/ Summarized: regular websites host more malware than sites dedicated to pornography.
Your cite is flawed. Or at least the conclusion you draw from it. The only conclusion that can be drawn is there is a greater number of infected non-smut websites than smut websites. This says nothing about the frequency of infections of smut websites vs. non-smut websites- in other words, it says nothing about the odds of infection given the fact that you are visiting a smut or non-smut website.
After spam, pornography is one of the largest demographics of internet traffic.
Does that still fail to leave a bad taste in your mouth?
It's bullshit for the same reason.
Also quite relevant is this site: http://www.yangsky.com/researches/physicallinguistics/PLUnde... Besides a "breast detector/nipple detector" he also created a couple of weird detectors like a "cowgirl sex position detector" among others. (The link is sfw, the individual detectors are not)
and there are lot of research papers if you are interested in implementing them : http://portal.acm.org/citation.cfm?id=1021261 http://www.springerlink.com/content/e774824834226tv2
It's a Python server side client for the cheap-but-paid PIFilter web service.
Hmm.. I guess it depends on what your definition of nudity is.
Also what about people with unmaintained pubic hair?
Other tools Im familiarmwith do body part detection, which words pretty well, and generates few false positives (that said, when they do occur they tend to be amusing - a pastrami sandwich , for example).
If the test is based on skin colours, what happens with different tones?
Perhaps a better approach would be attempting to characterize certain features, like nipples/areola or buttocks. That definitely sounds like it would require more intensive processing, though, and be highly dependent on the angle of the image.
I don't really know much about image processing and computational analysis, I just like to pretend I know things.
Thank you. I really want Firefox to be even slower running unnecessary JavaScript.
pretty useful concept, i must say...