Speech translation method and device, intelligent wearable device and storage medium
By using bone conduction sensors (VPU) and microphones in smart wearable devices for speech separation and relevance detection, the difficulty of translation when two parties are speaking simultaneously is solved, enabling barrier-free communication and improving the intelligence of the device.
CN122157693APending Publication Date: 2026-06-05GOERTEK INC
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GOERTEK INC
- Filing Date
- 2024-12-05
- Publication Date
- 2026-06-05
AI Technical Summary
Technical Problem
When both parties are speaking at the same time, smart wearable devices cannot effectively translate the speech, which affects barrier-free communication.
Method used
Using a bone conduction sensor VPU and a microphone, speech signals are acquired in real time and two-way speech detection is performed. Through speech separation and relevance detection, the speech signals of the speakers are identified and translated.
Benefits of technology
It enables seamless communication when both parties speak simultaneously, improves the intelligence of smart wearable devices, and ensures the accuracy and timeliness of translation.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122157693A_ABST
Abstract
The application discloses a speech translation method and device, an intelligent wearable device and a storage medium, and relates to the technical field of speech processing. The method comprises the following steps: collecting a first microphone signal and a VPU signal in real time, and performing double-speaking detection processing on the first microphone signal; if the detection result of the double-speaking detection processing is that the first microphone signal comprises speech signals of at least two users in the same time step, performing speech separation processing on the first microphone signal to obtain at least two groups of second microphone signals; for each group of second microphone signals, detecting the correlation between the second microphone signal and the VPU signal in the same time step; if the correlation is less than a preset correlation threshold, determining that the second microphone signal is a speaker speech signal, and outputting the speaker speech signal after translation. The application realizes barrier-free communication when both parties of a conversation speak at the same time, so that the intelligence of the intelligent wearable device is improved.
Need to check novelty before this filing date? Find Prior Art