TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

An Algorithm for Nudity Detection [pdf]

15 pointsby symisc_develover 11 years ago

10 comments

chewxyover 11 years ago
When I was in the advertising business, one of the core products I created was a brandsafety product - basically preventing advertisers from advertising on dodgy sites.<p>I messed around with algorithms that detected nudity (because big brand advertisers don&#x27;t want their ads showing up on porn sites). One of the more interesting and simple-to-use one is actually a simple averaging of the images across multiple samples. That one was easy to implement and has relatively good results.<p>In the end though, I ended up not using it because text clustering algorithms worked better in classifying content.
nathancahillover 11 years ago
&quot;The training set for the skin filter consisted of 1,182,608 manually labeled skin pixels and 10,471,553 manually labeled non-skin pixels while the testing set consisted of 2,303,824 manually labeled skin pixels and 24,285,952 manually labeled non-skin pixels.&quot;<p>That&#x27;s a lot of pixels to manually label.
评论 #7280565 未加载
评论 #7281033 未加载
adam-fover 11 years ago
This could only be a very rough first-pass on detection. Bathing suits can be very skimpy without being fully nude.<p>And social context plays a large role, for instance distinguishing between a fat male&#x27;s nipples and a small-chested female&#x27;s nipples would be impossible without analyzing a lot more than skin color.<p><a href="http://i.imgur.com/sb6Iw.jpg" rel="nofollow">http:&#x2F;&#x2F;i.imgur.com&#x2F;sb6Iw.jpg</a>
dclowd9901over 11 years ago
Seems like a not very scalable approach to the problem. I would think if you wanted to capture <i>all</i> nudity (including monochromatic or illustrated), you would instead go at the problem from the angle of <i>titillation</i>. You could even round up images that are not necessarily human based (fruit arranged provocatively, for instance).
评论 #7280544 未加载
ivan_ahover 11 years ago
direct link: <a href="http://onebit.us/x/i/814381733331796005.pdf" rel="nofollow">http:&#x2F;&#x2F;onebit.us&#x2F;x&#x2F;i&#x2F;814381733331796005.pdf</a>
saganusover 11 years ago
Uhmm... I think I&#x27;ve seen this somewhere else...
ganduGover 11 years ago
I wrote a really quick Python implementation of this a few months ago, in case someone is interested: github.com&#x2F;ParthGandhi&#x2F;nude.py
jplurover 11 years ago
I&#x27;m under the impression that Google&#x27;s machine learning is already filtering your personal G+ photos for nudity.
评论 #7280595 未加载
评论 #7280756 未加载
naderover 11 years ago
Does anybody know if there is an API that does this?
notastartupover 11 years ago
how do these nudity detection API work? Is there a crowdsourcing going underneath the hood? Are they using some clustering algorithm to detect a range of skin color (if 90%), it&#x27;s nude.
评论 #7280438 未加载
评论 #7280402 未加载
评论 #7280420 未加载