"TypoSwype: An Image Recognition Tool To Detect Typosquatting Attacks"

TypoSwype, an alternative tool for detecting typosquatting attacks based on image analysis, was recently developed by researchers at Ensign InfoSecurity, an end-to-end cybersecurity service provider based in Singapore. This tool, which was detailed in a pre-published paper on arXiv, uses advanced image recognition techniques to convert strings into images that take into account the location of letters on the keyboard. Typosquatting is the use of typographical errors to mislead users into visiting unwanted websites, according to David Yam, one of the study's researchers. Current anti-phishing techniques use string edit distance, which does not rely on keyboard character positions and thus is less accurate in detecting typos. In order to improve typosquatting detection, the researchers used image recognition techniques such as Convolutional Neural Networks (CNNs) and specific loss functions. TypoSwype is capable of capturing the distance between different characters on the keyboard by tracing lines between the buttons of consecutive characters on an imaginary keyboard. This helps reduce the errors in existing string edit distance metrics, which are methods that calculate how dissimilar two words or character sequences are from each other. In a series of tests, the researchers compared the performance of their typosquatting detection tool to that of the DLD algorithm, a widely used cybersecurity model. TypoSwype was found to be more reliable than DLD at detecting typosquatting while also accurately identifying the well-established and safe domains that attackers were attempting to copy or typosquat. This article continues to discuss the TypoSwype tool developed by researchers at Ensign InfoSecurity to detect typosquatting attacks. 

NU reports "TypoSwype: An Image Recognition Tool To Detect Typosquatting Attacks"