Image CAPTCHA Using Segmented Groups for Meaning Determination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image CAPTCHA systems face challenges in accurately determining the meaning of images, leading to lower accuracy compared to human interpretation, as they struggle to handle ambiguous text and multiple image meanings, and are vulnerable to test-wasting attacks by robots.
Innovation Solution
A method and server for providing an image CAPTCHA that includes a first group of images used to determine test passage and a second group of neutral images, where the test passage is determined solely by the first group, preventing exposure of image identifications and allowing for the automatic determination of image meaning based on user selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If conventional image search systems analyze text to interpret image meaning, then automated processing is achieved, but accuracy is lower than human interpretation due to ambiguous text and multiple image meanings
Solution Approach 1:
The patent introduces an intermediary process where human users act as mediators between the image and the automated system. Users select images from CAPTCHA options, and their selections are used to determine image meanings. This intermediary human judgment step bridges the gap between automated processing and accurate interpretation, allowing the system to leverage human cognitive abilities while maintaining automation benefits.
2Productivity
If robots attempt to bypass CAPTCHA by automatically solving tests, then account creation speed increases, but security is compromised due to test-wasting attacks
Solution Approach 1:
The patent applies preliminary action by pre-establishing a database of image meanings through human user selections before the security threat arises. When robots attempt test-wasting attacks, the system already has predetermined image interpretations to validate against, enabling rapid automated verification that prevents account creation abuse while maintaining security. The preliminary categorization of images by meaning allows the system to quickly determine whether selected images match expected patterns.
Solution Approach 2:
The system implements feedback mechanisms where user selections are continuously analyzed to refine and update image meaning determinations. This feedback loop allows the system to learn from human judgments and improve its automated verification capabilities over time, making it progressively more difficult for robots to bypass the CAPTCHA while maintaining user-friendly operation.
3Ease of manufacture
If text-based CAPTCHA is used, then automated processing is easier, but accuracy in determining image meaning is lower compared to direct human interpretation
Solution Approach 1:
The patent segments the image interpretation task into two distinct parts: (1) human users perform the complex interpretation task by selecting images from CAPTCHA options, and (2) the automated system processes these selections to determine meanings. This segmentation allows each component to operate at its optimal capability level - humans at interpretation and machines at processing - thereby achieving both accuracy and automation.
Data Source
AI summary
A method of providing an image Completely Automatic Public Turing test to tell Computers and Humans Apart (CAPTCHA), and a server thereof. The method includes: outputting an image CAPTCHA including a plurality of images including a determination object image; receiving an image selection from the image CAPTCHA; determining whether a test of the image CAPTCHA is passed, based on the received image selection; and determining what the determination object image means. The images include a first group of images used to determine whether an accessing party passes the image CAPTCHA test and a second group of images unused to determine whether the accessing party passes the image CAPTCHA test. The second group includes the determination object image. The determining of the pass of the image CAPTCHA test is performed by selecting an image representing a correct or incorrect answer from the first group.