A three-dimensional multi-modal interaction input determination method and related products
By constructing a multimodal interaction method that combines 3D motion sensing and voice input, and combining it with the context information of the interactive interface, a clear set of interactive command fields is generated. This solves the problem that the execution results in existing 3D interactive tools do not match the user's expectations, and improves the user experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING REALSENSE BYTE TECHNOLOGY CO LTD
- Filing Date
- 2026-05-27
- Publication Date
- 2026-07-21
AI Technical Summary
Existing 3D interactive tools suffer from fragmented and uncommon interaction modes in virtual character creation, partial character editing, and game scene construction. This leads to results that do not meet user expectations, resulting in semantic confusion of input and deviations in command recognition, thus reducing the user experience.
By constructing a set of spatial interaction parameters based on three-dimensional haptic input and a set of semantic intents based on voice input, and combining the contextual representation information of the current interactive interface, the interpretation constraints are determined, semantic rule mapping and legality filtering are performed, an initial set of control and spatial operation instruction fields is generated, and confidence calculation and structuring are performed to obtain the target set of interaction instruction fields.
It effectively reduces the deviation between the execution result and the user's expectations, improves the user experience, and fully restores the user's operation requirements through multimodal information collection, avoiding misidentification and invalid commands, and ensuring that the system's recognized intent is consistent with the user's intent.
Smart Images

Figure CN122431535A_ABST