Partial CAPTCHA Image Display for OCR and Solver Resistance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
CAPTCHAs are ineffective in preventing automated abuse as they can be easily broken by Optical Character Recognition (OCR) systems and third-party human solvers, leading to disruptions in user experience and increased costs for websites.
Innovation Solution
A method and system that provide an automated public Turing test by displaying only a predetermined portion of a CAPTCHA image to users, using client-side code to hide parts of the image and perform User-Input Analysis to detect fraud, including checks for relay attempts and device fingerprinting to verify user identity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If CAPTCHA challenges are made more difficult (blurred, skewed, overlapping characters) to prevent OCR-based breaking, then OCR resistance is improved, but user experience deteriorates and human solving becomes feasible
Solution Approach 1:
The CAPTCHA image is divided into multiple segments or regions, with only a predetermined portion (e.g., specific characters or area) being displayed or accessible to the user. This segmentation prevents OCR systems from analyzing the entire image while still presenting a solvable portion to human users, thereby maintaining OCR resistance without excessively complicating the challenge for legitimate users.
Solution Approach 2:
Only the necessary or predetermined portion of the CAPTCHA image is extracted and displayed to the user, rather than showing the complete image. This extraction approach limits the information available to OCR systems while providing sufficient information for human users to solve the challenge, thus improving OCR resistance without unduly burdening legitimate users.
2Reliability
If CAPTCHA challenges are made more difficult to prevent OCR-based breaking, then security against automation is improved, but third-party human solving becomes more effective
Solution Approach 1:
By segmenting the CAPTCHA image and displaying only a predetermined portion, the system creates a challenge that is sufficiently difficult for automated OCR systems but remains solvable for human users. This segmentation strategy effectively raises the bar for automation while maintaining feasibility for legitimate human users, thereby reducing reliance on expensive third-party solving services.
Solution Approach 2:
The difficulty parameters of the CAPTCHA (such as blur level, skew angle, character overlap) are adjusted to optimize the balance between OCR resistance and human solvability. By carefully tuning these parameters, the system achieves sufficient difficulty to deter automated attacks while maintaining ease of use for legitimate users, thus reducing the effectiveness of third-party solving operations.
3Ease of operation
If complete CAPTCHA images are displayed to users, then user experience is maintained, but OCR systems can successfully decipher the challenges
Solution Approach 1:
The system extracts and displays only a predetermined portion of the complete CAPTCHA image to the user. This extraction approach maintains adequate user experience by providing sufficient visual information for solving, while simultaneously reducing OCR resistance by limiting the total information available to automated systems.
Solution Approach 2:
The CAPTCHA image is segmented into multiple parts, with only the relevant predetermined portion being displayed. This segmentation allows the system to present a solvable challenge to users while withholding enough information to prevent successful OCR analysis, thus balancing user experience with security.
Data Source
AI summary
A method and system of administering a CAPTCHA challenge comprising: providing a CAPTCHA challenge to the client computer, the CAPTCHA challenge comprising an image and browser code for displaying the image, the image comprising a character string in which a first sub string smaller than said character string corresponds to a solution to the CAPTCHA challenge and a second substring smaller than said character string corresponds to extra characters not part of the solution, displaying the image on the client computer wherein only the first substring is visible in the browser, receiving a response from the client computer, comparing the response with the second substring, and determining, if the response includes the second substring, that the client computer used a CAPTCHA bot or relay.