13 {{ upvoteCount | shortNum }}
13 {{ upvoteCount | shortNum }}

Where can I find someone to make a script to decode this captcha?

by davidwa
10 replies
I'm looking for someone to create a script that will work on Linux to decode a captcha:

https://s27.postimg.org/gi3zcb483/DXB.png?noredir=1

It's made difficult due to the close together letters.

I hired one person from Freelancer to create a script, but they disappeared. There doesn't seem to be anyone on there who has the skills to do this.

Does anyone know where I could post to find someone who might be able to do this?

A human captacha solving service is not an option due to volume.
#captcha #decode #find #make #script
  • Profile picture of the author judaculla
    Youd save waaay more money by just integrating with a captcha service like deathbycaptcha.

    A custom 'script' would involve building a visual recognition system, and would likely extend far beyond the application of any freely available libraries. Even then, youd have to train it on the types of captchas youre dealing with.

    Even enterprise level software integrates with 3rd party services. The implications of a custom build for something like that are disheartening, and you're freelancer likely realized that.
    {{ DiscussionBoard.errors[10972125].message }}
    • Profile picture of the author davidwa
      I've made a number of scripts myself to decode captchas. Relatively simple in many cases. It's relatively cheap and easy and can be done with PHP/Python + the open source OCR, Tessercat.

      But I don't have the time to do it, nor is this something I am specialised.

      I used to know a guy who specialised in them, but have lost contact.
      {{ DiscussionBoard.errors[10972163].message }}
  • Profile picture of the author yukon
    Banned
    You would need a couple of things:
    • Split the image up based on average pixel width for each character.
    • OCR API for converting each image character into plain text
    {{ DiscussionBoard.errors[10972136].message }}
  • Profile picture of the author DavidAllenNeron
    Black Hat Tactics Much?

    I really wish the internet didn't come to this ... but it is, what it is, eh?
    Signature
    {{ DiscussionBoard.errors[10972138].message }}
    • Profile picture of the author davidwa
      It's not for malicious purposes, i.e. posting spam. Though I guess the site owner would prefer I not scrape their database for personal use (not for republishing).

      Another day in the web sewer...
      {{ DiscussionBoard.errors[10972164].message }}
      • Profile picture of the author outscrape
        And remember, big big companies are blowing through captchas with machine learning, using proxies, etc. I don't want to downplay it entirely but there are big big industries built on scraping, and companies in industries built on scraping, etc. Hope that doesn't sound patronizing, just been doing a lot of web scraping research lately and realized it's not just blackhatters.

        I'd check upwork. But it's probably a $200 job.

        Captcha services are about $.02/captcha, so you'd probably need to solve about 10k of those bad boys before it's worth getting a custom solution.
        Signature
        Tired of methods that die out? Learn web scraping secrets that work in any business.
        {{ DiscussionBoard.errors[10972277].message }}
      • Profile picture of the author David Beroff
        Originally Posted by outscrape View Post

        Captcha services are about $.02/captcha....
        I'm definitely not in this space, but the last time I looked at this for a thread here, pricing was more like US$0.60 per thousand.

        Originally Posted by davidwa View Post

        It's not for malicious purposes, i.e. posting spam. Though I guess the site owner would prefer I not scrape their database for personal use (not for republishing).
        What you intend to do with the data is irrelevant. They're protecting their site with CAPTCHA, and you're trying to circumvent this. I'd call that malicious.
        Signature
        Put MY voice on YOUR video: AwesomeAmericanAudio.com
        {{ DiscussionBoard.errors[10974076].message }}
  • Profile picture of the author 3ym3jtmmk
    Banned
    [DELETED]
    {{ DiscussionBoard.errors[10974091].message }}
    • Profile picture of the author David Beroff
      Originally Posted by outscrape View Post

      And remember, big big companies are blowing through captchas with machine learning, using proxies, etc. I don't want to downplay it entirely but there are big big industries built on scraping, and companies in industries built on scraping, etc. Hope that doesn't sound patronizing, just been doing a lot of web scraping research lately and realized it's not just blackhatters.
      Originally Posted by 3ym3jtmmk View Post

      enormous huge organizations are blowing through captchas with machine learning, utilizing intermediaries, and so on. I would prefer not to minimize it completely however there are huge businesses based on scratching, and organizations in enterprises based on scratching, and so on. Trust that doesn't sound belittling, simply been doing a great deal of web scratching research recently and understood it's not quite recently blackhatters.
      Which spinner software did you use to generate that?
      Signature
      Put MY voice on YOUR video: AwesomeAmericanAudio.com
      {{ DiscussionBoard.errors[10974099].message }}
  • Profile picture of the author AboutTown
    This does not look like a very complicated captcha. I would assume that the curve/shape of the letters varies as well as the colour of the letters. Maybe even the font? More samples would be needed to assess these things.

    Taking a look at the page that produces these captchas could help to determine what software/technology is being used to make them. That would hep a lot to come up with a solution!

    The most successful captchas use background noise to confuse OCR scanners. I don't think bunching up the letters and curving them is enough to beat an automated solver.
    {{ DiscussionBoard.errors[10984204].message }}

Trending Topics