Google Using ReCAPTCHA To Decode Street Addresses 104
smolloy writes "Apparently some users of reCAPTCHA have recently begun seeing photographs appear in their CAPTCHA puzzles — photos that look very much like zoomed in house numbers taken from Google Streetview. It appears that Google has decided to put the reCAPTCHA system to help clean up Google streetview images, and 'according to a Google spokesperson, the system isn't limited to street addresses, but also involves street names and even traffic signs.' A large collection of these has appeared on the Blackhatworld website."
Take off your tin foil hat (Score:4, Insightful)
This is an incredibly fascinating and great use of the technology.
Re: (Score:2)
I'm guessing you've never done a copy-and-paste on, say, Google Books because the OCRed text quite frequently contains typos, random inserted spaces and completely wrong words. And since reCapatcha is used to supplement the OCR on Google Books, it would appear they aren't as smart as you would like them to seem.
Re: (Score:2, Insightful)
It's mostly the fault of 4chan.
Ever since Re-Captcha was implmented there, most of the RC results are
'(Checkword) Nigger'
Re: (Score:2)
I'm guessing that these books haven't been subjected to recaptcha yet. Considering the number of books that google has scanned, this could take a while.
Re: (Score:2)
Re: (Score:1)
It is not really unpaid work. You are "working" so that you can use the "free" recaptcha service.
Remember, if a something is free, you are the product being sold.
Eyebleed site (Score:1)
Wow that site is so terrible looking that it makes Geocities and myspace look decent. The only thing it's missing is cosmic cursors.
Re:Eyebleed site (Score:4, Informative)
Wow that site is so terrible looking that it makes Geocities and myspace look decent. The only thing it's missing is cosmic cursors.
Yeah, Techcrunch is really ugly isn't it.
Re: (Score:1)
Baziiiing!
Re: (Score:3)
Baziiiinga!
Fixed it for you.
If I just type out the necessary word... (Score:3)
Re: (Score:2)
Re:If I just type out the necessary word... (Score:5, Insightful)
Great. You know what they were previously? OCR for things like libraries.
I think your own answer to them describes what you are.
Would make for some interesting kids' books (Score:3)
I do not like FUCK Sam I am
Re: (Score:2)
... and you just put the idea of "Green Eggs and Ham" censored, just like The Count [youtube.com]. Disclaimer: I am not responsible for any injury occurring as a result of laughing at this video.
Re: (Score:2)
You are a loser.
Re: (Score:2)
Re: (Score:2)
Can we all agree on a word for the addresses just to have some fun with google?
Actually, words instead of numbers could be an issue already. My parents' house does not have a number anywhere. The house has a visible name instead, and that's what is used in letters addressed to them (including government letters): house-name, street-name, etc. Some houses on their street have numbers, but most just have names, and the house names are nothing to do with the names of the occupants. BTW that particular first world country does not have any postal codes, either.
Re: (Score:2)
Ireland seems most likely.
Bingo, AC got it. I think almost every other EU country has postal codes - at least they do where I live. Incidentally, I've had to complain to more than one web-shop in the EU since they have the postal code as a required part of the address. So when ordering a gift for a parent, I have to put some bogus crap down (e.g. repeat the town name) as their "post code".
Re: (Score:2)
In the USA when e911 service is introduced into an area each street is named and each house numbered in the maps e911 uses, I assume these are official postal addresses as well. It is a good idea to have your house number visible for emergency services to find you.
Re: (Score:1)
Re: (Score:2)
If they're using this as a way to identify the street numbers, then I would assume that they're randomly matching the numbers with different words and seeing if they can get several matches to the same numbers. I would guess that they're also comparing the results to attempts at automated OCR. It would be difficult to bomb.
Re: (Score:2)
This is actually kind of frightening... (Score:1)
Could just be me being paranoid, but this sounds like something out of a science fiction book. Whoever had the idea to do this, I have to admit, was really using their head though.
Re:This is actually kind of frightening... (Score:5, Insightful)
Re:This is actually kind of frightening... (Score:5, Funny)
Oh shit
http://www.whitepages.com/ [whitepages.com]
Re: (Score:2)
Interesting. It doesn't show the street numbers. That is bloody pointless.
Looks like out whitepages is far more useful [whitepages.com.au]
Re: (Score:2)
Click on the name and you get their street number as well, sometimes. That's what he meant by "most of the time".
But yes, us civilized (tongue in cheek) countries have way better white pages.
Re: (Score:1)
Seriously, there are privacy invasions out there that actually matter. Making public information more public is not a privacy invasion, and it makes for a "boy who cried wolf" appearance.
Re: (Score:2)
Re: (Score:2)
Theoretically, there is enough information in the actual font of the text, the frame that the text is on, and the material and texture of the wall to identify that location uniquely.
If you have seen that sign before, you would be able to recognise that location. USA street names tend to be white text on small blue rectangular signs at 90 degrees to each other, and on posts. London street names tend to be white rectangular plates mounted on walls along with the postcode at the bottom. In Scotland, thr street
Re: (Score:2)
What is your point? Who cares where a street sign is from? None of them are private information. And you're wrong about US street signs, they vary regionally but the majority tend to be white on green. And we were talking about house numbers anyway, or so i thought. How is any of what you wrote relevant?
Re: (Score:2)
You mean your public information is public? shocking.
"Could just be me being paranoid, but this sounds like something out of a science fiction book."
Warp Drive teleporters, FTL, light sabers and robots are all in Science fiction. Why do you think things in science fiction are bad?
Re: (Score:2)
I just raised a question, I even stated that I "Could just be me being paranoid". When did
Re: (Score:2, Interesting)
They give you two words to solve. One is an old, known word and the other is a new, unknown word. You have no way to tell which is which. To pass the CAPTCHA, you need to answer both and get the known one correct. Eventually entries can go from unknown to known when enough people provide the same answer.
Re:How does ReCAPTCHA "solve" new images? (Score:4, Informative)
They give you 2 words, one is an already solved known value, and the other is an unknown word.
if you get the first word correct, they take the value from your second word and add it to the "possible solutions" list.
After 2000 or so people have solved the word, they examine the results for a statistically unique answer. If there is not outlier, (say 65% have the same answer) it goes back into the unknown pile.
Once they find a statistically significant answer, it's considered "solved" and is used as one of the initial validation words.
Rinse, repeat.
I seem to have missed something... (Score:5, Informative)
I have read the quote from Google about what they are doing several times, and I don't see what everyone else sees. It appears to me that they are using the already known street names and numbers as possible ReCAPTCHA images. What they are NOT doing is using the results given by people to define what the image says. The point of the experiment is to determine whether these images are sufficient to separate people from web-bots. I imagine that they will look at the number of 'wrong' answers from both sides of the test, and see if bots are able to parse the street view images significantly more often than the standard test images.
So... can anyone point to something in the Google quote to show me where I went wrong? From TFA, here is the quote:
We’re currently running an experiment in which characters from Street View images are appearing in CAPTCHAs. We often extract data such as street names and traffic signs from Street View imagery to improve Google Maps with useful information like business addresses and locations. Based on the data and results of these reCaptcha tests, we’ll determine if using imagery might also be an effective way to further refine our tools for fighting machine and bot-related abuse online.
Re: (Score:3)
Getting around reCAPTCHA logins is usually easy. Just correctly type the easy to read word, and an approximation of the number of characters in the hard to read one. You don't even have to be close.
Google could have a few thousand house numbers they already know (their own recognition system is probably capable of this), and they can swap these in as well as a hard to read scanned word from a book, and you could never be sure which one was the reCAPTCHA and which was the CAPTCHA.
Re: (Score:2)
Yes, I understand this. I understand that they can look for most common answers among correct control responses, and crowd source the OCR of difficult street view images. My point is that is not what the experiment is doing. The point of the experiment is to determine if these images are as effective as the current images used in the tests. For the purposes of that experiment, it would be much easier (and probably more scientifically accurate) to use images where the correct answer is already known. As
Re: (Score:2)
Who is asking who to panic? What are you going to panic about?
Um, what? That's exactly what they're doing. (Score:2, Informative)
What they are NOT doing is using the results given by people to define what the image says.
Um, no, that's exactly what ReCaptcha is for! The standard ReCaptcha images are all from old books that were scanned in (and presumably had trouble being OCRed with high confidence), and Google used ReCaptcha to "read" the words.
For heaven's sake, ReCaptcha's MOTTO is: "reCAPTCHA: Stop Spam, Read Books"
I read how it works. Multiple users are shown the same image, and once a few people have identified a given image as the same word, it's treated as the "correct" answer, and then later users have to match t
Re:Um, what? That's exactly what they're doing. (Score:5, Interesting)
It's quite noticeable if you use a site which relies heavily on recaptchas. For example, when you get a word which has old english S [wikipedia.org] which looks like a modern small case F, you're much better off claiming it's an F instead of giving the correct answer.
Are people actually annoyed at this? (Score:1)
They already list the number of the house on maps.
Re: (Score:1)
My understand of ReCAPTCHA is that it's to help translate books for libraries. Google has distorted that by using it to improve it's own databases. I personally don't have ReCAPTCHA on my website, but if I did I would be completely pissed off. Google is a for-profit company and can pay to do user studies to see how well people can read images. I'm willing to donate my time/reading ability to random libraries, not Google.
Re:Are people actually annoyed at this? (Score:5, Insightful)
Oh, climb down off that ledge before you get hurt.
reCAPTCHA is for what ever you want to use it for, Its simply a technique for crowdsourcing guesses.
In my estimation, Google maps and street view is one of the great accomplishments of our time, easily worth every penny Google monetizes out of it.
Re: (Score:2)
The picture of my house has a funny join; two photos that span my house. The join is in the middle of my burgular siren and it makes it stand out. Brilliant.
Thank Goodness! (Score:2)
I'm glad something is being done I can't recall how many times I've looked up a street address to find Google maps reporting it as being 4 or 5 blocks away (on average) from where it actually is.
At this point I only solve the generated word. (Score:2)
Back when reCaptcha showed two words that you could find in the dictionary, black on white I had no problem with it, it seemed like a good idea and you might be contributing to digitizing a book or something.
But now you just get randomly generated characters with a zigzag going through the middle and blobs that invert it and it's hard to tell if this one letter is an 'i' or an 'r' or a 't'.
So I don't even bother looking at the real word and just solve the generated one.
So they aren't hiding them anymore? (Score:2)
So many addresses has been fuzzy that I could that could only be a strange design choice.
Unclean by design? (Score:2)
I thought text in Streetview was blurred out by design in the same way that faces were-- automatically and for security reasons (read: so Google doesn't get sued by crazy OMG I'M ON TEH INTERNET people).
I'd actually prefer if they un-blurred all street numbers and signs. It's fine to rely on Map's street number location when you're in a huge city, and the difference between 123 fake street and 125 fake street is ten feet or so. But last time I planned a ro
Re:I'm a Microsoft whore (Score:5, Insightful)
Yeah because those street number designed to tell everyone passing by what number the house is on the street are meant to be private.
Re: (Score:1)
Don't feed the trolls... (but you're right, though)
Re:I'm a Microsoft whore (Score:5, Informative)
Yet Google would have to know what the address numbers really was in order to validate the reCAPTCHA, so that can hardly be why they are doing it. They don't need to crowd source an answer that they already know.
No they don't. They also add an altered text image alongside the picture (which presumably they generated), and can use that to validate the CAPTCHA. The street number can be validated by numerical probability (if 70% of them say it is "257", and the numbers "2,5,7" appear frequently in the rest, it is probably "257") even if they don't already know what it is.
Re: (Score:2)
Plus, street numbers in the US typically go odd/even on either side of the street, so they can extrapolate most of the time.
Re:I'm a Microsoft whore (Score:5, Informative)
Interesting (Score:1, Insightful)
I only have about a 60% success rate on those swirly semi-inverted ones. My wife's friend's decaptcha software does a much better job than I do with its 79% success rate. I had wondered that as they get harder to read that the day was almost here when only machines would have the ability to decode captchas and prove that they were human.
Re: (Score:1)
ReCaptcha will accept any sequence of symbols for the unknown word. The most telling sign that a word is unknown is that, out of the two, it is the one that is ACTUALLY A WORD. Other signs are non-standard fonts, scanning distortions, non-Latin symbols, and punctuation marks.
Furthermore, there is a 1-chacter fault tolerance for the sequence of letters used as the part of the ReCaptcha to actually check if you pass or fail or not.
Re: (Score:1, Funny)
ReCaptcha will accept any sequence of symbols for the unknown word.
Wait - you can type something other than "nigger" for the unknown word?
One of these days I'm going to do that when someone's looking over my shoulder and get a serious WTF from them.
To whoever modded the parent at -1, pay attention.
Out of the two images you are presented, one is known, the other is unknown. When a large enough number of people have entered the same answer for the unknown image, it gets moved to the 'known' list with that particular answer.
So on some places like 4chan, there has been a large effort to get as many people as possible to answer the unknown image with the word 'nigger'. If enough people do it on a single unknown image, it will get added to the pool with the
Re: (Score:1)
Re: (Score:1)
Fun.
I didn't know about the nigger thing, but I've always submitted nonsense for the book one.
Re: (Score:2)
Yet Google would have to know what the address numbers really was in order to validate the reCAPTCHA, so that can hardly be why they are doing it. They don't need to crowd source an answer that they already know.
Doubtful. They post two images. One they know and one they don't. They use the data for the one they don't, combine it with data from 1000s of other people who have also solved that captcha to get an accurate picture of what that particular number is. They use the one they know to validate the recaptcha data and verify you're human...
Re:I'm a Microsoft whore (Score:4, Informative)
One of the two words is considered solved, and is the actual captcha, the second word is using you as an ocr.
After enough people provide the same solution for the second word, it goes into the solved category and is used for validation.
They don't have to pay people to validate the addresses, we're doing it for free.
Re: (Score:2)
But unless Google is paying a zillion people to validate these images visually,
That is EXACTLY what Google is doing. And the payment is access to the site the reCAPTCHA is protecting.
Be a Roman harlot instead! (Score:5, Funny)
Re: (Score:2)
Re:Be a Roman harlot instead! (Score:4, Funny)
And put your house number in Roman Numerals. Nothing like living in number CLXXIV to screw up the recaptcha.
Not to mention the postal service! Damn snooty mailmen with their eagle-logo cars and fancy uniforms... Now I know how to get back at them.
Re: (Score:2)
And put your house number in Roman Numerals. Nothing like living in number CLXXIV to screw up the recaptcha. Anyone answering with 174 is likely counted as wrong...
And then hope that the ambulance driver trying to find your place is skilled in the Roman ways.
Re: (Score:1)
Re: (Score:1)
Awesome. More invasion of privacy. Fuck Google.
bonch
What makes this more of an invasion of privacy than whatever they used to do to find house numbers? I assume they used some combination of databases, OCR, and paying someone to do it.
I'm surprised that this is a big help to them - if they can identify that something on a house is the house number (as opposed to a shadow or some home design pattern), it's surprising that they can't identify the number itself. It seems like there's going to be relatively few instances where something is identifiable as a hous
Re: (Score:3)
Different angles make it hard to be sure you have the number right. If you look at a street photo like a book you're going to OCR, you have first the layout detection, then identify the image part and the text part. Solving this problem would be similar to identifying where the page number is, to be eliminated from the text.
Taking a laser measurement, un-warping the photo, and then doing traditional OCR would be awesome, if they had the forethought to include the laser part in their vast collection, but t