A phishing attack on your organization can have a huge impact on the whole workforce. It is important to detect and respond to phishing attacks quickly to minimize the spread of the threat.
Anti Phishing With Detection Powered By Visual-AI
When a user enters sensitive information into an online form, threat actors can steal that information. Phishing detection and response tools can save administrators time and help them respond to threats more efficiently.
Several techniques have been proposed for phishing website detection. These methods can detect suspicious URLs through visual, text or domain features. However, these methods have limitations. They are slow and cannot handle the changes that occur on phishing websites.
Visual similarity-based approaches compare the visual contents of a suspected website with the contents of a legitimate website. However, this type of approach can produce false positive results when the visual contents of the suspected webpage change.
In order to address the shortcomings of visual similarity-based techniques, some researchers have proposed methods based on multi-level features. Such methods use multiple classifiers to improve generalization.
In addition, these authors consider web image properties, such as color and coordinate, to generate signatures. Finally, they compute the distance of webpage image signatures to authentic and suspected webpages using EMD.
Another method to detect phishing websites uses an apriori algorithm. This technique identifies the phishing domain in 93 percent of the data. However, this method is more complicated and requires longer training time.
To improve the accuracy of the CNN model, the authors used several classifiers to improve the model’s parameters.
