Skip to main content

    Stanford researchers crack video CAPTCHA

    Stanford researchers crack video CAPTCHA

    /

    After solving text and audio CAPTCHAs, researchers at Stanford University have successfully cracked video CAPTCHA, and have offered suggestions to improve the security measure.

    Share this story

    video captcha owned
    video captcha owned

    If you think CAPTCHAs are keeping spam bots at bay on the internet, think again. The same Stanford team that successfully cracked audio and text versions of the security feature turned their sights to video CAPTCHAs, and the results were similar. Video CAPTCHAs try to fend off bots by adding animation to the jumble of characters used to prove you're a human on the internet, but by modifying the Stanford-developed Decaptcha software and adding in optical flow algorithms, the researchers managed to demonstrate that over 90 percent of them can be solved by a computer. According to the security feature's creator, NuCaptcha, the video version was supposedly the best method for weeding out bots from websites.

    The researchers shared their results with NuCaptcha months ago, but are just now making their findings public. The team suggested ways to improve the system, including making animation patterns more erratic, tightening the letter groupings, and adding more decoy objects in the animation. NuCaptcha said it has worked in the suggestions, but as long CAPTCHA remains a computer vs. computer affair, we're sure it will eventually be cracked again. If you want to read about the gritty details of how video CAPTCHA was cracked, and a response from NuCaptcha, check the source links below.