面向三维设计引擎的模型特征编码与检索管理方法及系统
By separating fields and generating stable features, the problem of structure tree encoding in 3D design engines being sensitive to import order and number of instances is solved, achieving consistency in structural representation and retrieval accuracy under different conditions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GUANGZHOU GRAVITATIONAL WAVE INFORMATION TECH CO LTD
- Filing Date
- 2026-04-16
- Publication Date
- 2026-07-17
AI Technical Summary
In cloud-based model management and retrieval scenarios for 3D design engines, existing structure tree encoding is overly sensitive to fluctuations in import order, naming conventions, and the number of instances, resulting in incomparable encoding results and affecting retrieval accuracy.
By separating fields, generating stable features, performing collision detection and context refinement, a repeatable structural input is generated. When a version chain exists, a stability self-check is performed to generate structural fingerprints and structural vectors, suppressing the interference of non-essential changes on the encoding.
It improves the consistency of structural representation and retrieval accuracy under different versions, sources or parsing conditions, and reduces the impact of import order and instance number fluctuations on encoding.
Smart Images
Figure CN122152822B_ABST