A dynamic man-machine verification method based on access path modeling
By establishing an access path behavior feature model and calculating multiple abnormal indices for multi-stage risk assessment, the problem of difficulty in identifying abnormal interface access behavior in existing technologies is solved, and dynamic human-machine verification and security protection for application systems are realized.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XIAMEN KUAIKUAI NETWORK TECH CO LTD
- Filing Date
- 2026-04-01
- Publication Date
- 2026-06-23
AI Technical Summary
Existing technologies struggle to accurately identify and control abnormal application system interface access behavior, especially automated script batch calls and malicious crawlers, leading to security risks such as system resource consumption and data leakage. Furthermore, they lack multi-stage risk assessment and step-by-step verification mechanisms.
By collecting interface access logs, session behavior data, and business object processing data, an access path behavior feature model is established, and access path anomaly index, behavior rhythm deviation index, and business semantic conflict index are calculated to conduct multi-stage risk assessment and dynamic human-machine verification.
It enables comprehensive expression and quantitative analysis of user access path structure, time rhythm and business semantics, improves the accuracy of abnormal access behavior identification and system security protection capabilities, and enables dynamic intervention and closed-loop verification in the early stage of risks.
Smart Images

Figure CN122046323B_ABST
Abstract
Citation Information
Patent Citations
Cloud desktop security access control method based on zero-trust architecture
CN120750599A
Dynamic access blocking method based on zero trust
CN121217444A