Phishing Detection via URL Glyph Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Phishing attacks deceive users by creating bogus websites with URLs that resemble legitimate ones, using similar-looking characters or characters from different alphabets, making it difficult for users to distinguish between genuine and fraudulent sites.
Innovation Solution
A method involving image analysis and machine learning algorithms to compare glyphs of characters in URLs, calculate encoded distances, and check against thresholds to identify potential phishing sites, utilizing a domain name whitelist and edit distances to alert users of suspicious URLs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If users rely on visual inspection of URLs to identify legitimate websites, then the ease of operation is improved, but the reliability is worsened due to similar-looking characters in phishing URLs
Solution Approach 1:
The patent replaces manual visual inspection of URLs with automated image analysis technology. The system captures screenshots of web pages, extracts URL text, renders it as an image, and uses image recognition algorithms to analyze the visual characteristics of characters. This substitution of mechanical human inspection with automated optical analysis resolves the contradiction by providing both ease of operation (automatic analysis) and improved reliability (detection of visual similarities that humans might miss).
Solution Approach 2:
The patent introduces an intermediary image analysis system that acts as a mediator between the user and the URL verification process. Instead of directly comparing text strings, the system converts URLs to visual images and analyzes them through multiple parameters (character width, height, pixel density, font characteristics). This intermediary approach enables reliable detection of phishing URLs while maintaining ease of operation for users.
2Device complexity
If traditional text-based URL comparison methods are used, then the device complexity is reduced, but the measurement precision is worsened due to inability to detect visual similarities
Solution Approach 1:
The patent transitions from one-dimensional text string comparison to two-dimensional image analysis. By rendering URLs as images and analyzing visual characteristics such as character width, height, pixel density, and font properties, the system adds a spatial dimension to the detection process. This dimensional change enables precise detection of visually similar characters (like 'l' vs '1' or 'O' vs '0') that cannot be distinguished through traditional text comparison, thereby improving measurement precision while maintaining reasonable system complexity.
3Reliability
If automated image analysis is implemented for URL verification, then the reliability of phishing detection is improved, but the use of energy and computational resources is worsened
Solution Approach 1:
The patent applies partial action by selectively analyzing only critical visual features of URLs rather than performing complete image processing on entire web pages. The system focuses specifically on extracting and analyzing the URL text portion, comparing key character visual properties (width, height, pixel density) against known phishing patterns. This selective approach maintains high detection reliability while significantly reducing computational overhead and energy consumption compared to full-page image analysis.
Data Source
AI summary
A method for detecting potential phishing URLs includes extracting a URL from a document, analyzing the URL context, and comparing the URL to stored trusted URLs and stored known phishing URLs. The URL context includes anchor text and surrounding content associated with the URL. The method further includes generating a phishing alert based on the comparing and the analyzing.