Domain Name Visual Similarity Detection via OCR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for registering and acquiring certificates for domain names are inadequate in distinguishing between visually similar characters from different character sets, leading to potential phishing attacks where domain names appear identical to users but are technically different, and manual maintenance of confusingly similar characters is impractical due to the increasing variety of character sets.
Innovation Solution
The implementation of optical character recognition (OCR) to render domain names as images, extract text, and compare them against a blacklist of visually similar domain names, denying registration or certificate issuance if a match is found, using multiple fonts and tests to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual maintenance of confusingly similar characters is used, then accuracy in identifying visually similar domain names can be maintained, but the process becomes impractical due to the increasing variety of character sets
Solution Approach 1:
The patent replaces the manual mechanical process of maintaining lists of confusingly similar characters with an automated optical character recognition (OCR) system. The OCR technology automatically renders domain names as images, extracts text, and compares them to identify visually similar characters, eliminating the need for manual maintenance while handling the increasing variety of character sets efficiently
Solution Approach 2:
The system performs self-service by automatically identifying and updating lists of confusingly similar characters through OCR technology. The system continuously renders domain names as images, extracts text, compares characters, and updates the confusingly similar character lists without human intervention, adapting automatically to new character sets as they emerge
2Loss of information
If domain names are compared using Unicode representation, then technical differences between characters can be identified, but visually identical characters from different character sets cannot be distinguished
Solution Approach 1:
The patent adds a visual dimension to domain name comparison by rendering domain names as images and using OCR to extract text. This transforms the comparison from a purely character-code level (Unicode) to an image-processing level, enabling detection of visual similarity that transcends character set boundaries. The system compares both the rendered images and the extracted text to identify visually similar characters regardless of their Unicode representation
3Measurement precision
If multiple fonts and tests are used to enhance OCR accuracy, then identification of visually similar domain names improves, but system complexity increases
Solution Approach 1:
The patent segments the domain name comparison process into distinct modular components: rendering domain names as images using multiple fonts, performing OCR text extraction, comparing extracted text against original domain names, and maintaining lists of confusingly similar characters. This segmentation allows each component to be optimized independently while managing overall system complexity through clear separation of concerns
Data Source
AI summary
Techniques are disclosed for mitigating against registering a domain name that is confusingly similar to a pre-existing domain name, possibly for the purpose of fooling users. In embodiments, a domain name is presented for registration. The domain name is rendered as an image, and optical character recognition is performed on the image to extract the rendered text. This extracted text is compared against a list of domain names for which confusingly similar domain names cannot be registered, and when the extracted text matches a domain name in this list of domain names, registration of the domain name is denied.