文档解析与结构化处理方法、装置、设备及可读存储介质
By selecting the parsing mode based on document attributes and combining it with a multi-round inference model to process documents, the problems of poor compatibility and segmentation effect in multimodal document parsing are solved, and efficient and standardized document structuring processing is achieved.
CN122414166APending Publication Date: 2026-07-17THREE GORGES HI TECH INFORMATION TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- THREE GORGES HI TECH INFORMATION TECH CO LTD
- Filing Date
- 2026-06-17
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing technologies have poor compatibility in multimodal document parsing, poor performance in segmenting long documents, and rigid, coupled systems with poor scalability, making it difficult to meet the needs of practical applications.
Method used
The target parsing mode is selected based on document attributes. The document format is converted using parallel threading or a parsing engine. Text fragments are split and structured through a multi-round inference model, and the output format is combined with semantic analysis and business type constraints.
Benefits of technology
It improves the adaptability and standardization of document parsing, balances parsing efficiency and accuracy, and ensures the uniformity of converted documents and the accuracy of structured data.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122414166A_ABST
Abstract
一种文档解析与结构化处理方法、装置、设备及计算机可读存储介质。该方法中，通过获取待解析文档，并根据所述待解析文档的文档属性，确定目标解析模式；基于所述目标解析模式，将所述待解析文档转换为目标格式文本；基于多轮推理模型和所述目标格式文本，得到待解析文档的结构化数据，通过适配文档属性确定解析模式，将文档转换为统一目标格式，结合多轮推理模型获取结构化数据，兼顾解析效率与精度，提高文档解析适配性，保证文档解析规范性。
Need to check novelty before this filing date? Find Prior Art