A contract key fragment extraction method
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA ELECTRONICS CLOUD DIGITAL INTELLIGENCE TECH CO LTD
- Filing Date
- 2026-04-22
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies suffer from problems such as wasted computing resources, noise interference, insufficient semantic coherence, and poor traceability of evidence when extracting key information from long contracts. In particular, it is difficult to balance computing costs, noise control, and semantic integrity when dealing with mixed-format contracts.
The contract document is parsed page by page. Pages with fewer than a threshold of valid characters are marked as low-text pages and OCR is selectively performed on them. The physical page numbers of business keywords are recorded, and adjacent page numbers are expanded forward and backward from these as the center to build a set of business-related page numbers. After adding page number tags, the pages are concatenated and output.
Reduce computing costs, improve processing efficiency, enhance document adaptability, ensure semantic coherence and auditability of results, and improve the accuracy of key information extraction and system stability.
Smart Images
Figure CN122414162A_ABST