Voice endpoint detection method and device, computer equipment and storage medium

• Main Authors: WEI TAO, MA JUN, WANG SHAOJUN, TAN FENGYUN
• Format: Patent
• Language: Chinese; English
• Published: 22.11.2022
• Subjects: ACOUSTICS; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

The embodiment of the invention belongs to the technical field of audio processing in artificial intelligence, and relates to a voice endpoint detection method and device for vibration audio, computer equipment and a storage medium. In addition, the invention also relates to a block chain technology, and the target voice endpoint of the user can be stored in a block chain. According to the method, the effective data of the actual business scene is utilized for data enhancement, the problem of lack of non-human voice noise data such as mobile phone vibration is relieved, and therefore the pain point problem that the model learning effect is poor due to sparse vibration data in model training is solved; according to the method, the model can solve the problem of vibration misrecognition in a quiet and noisy background human voice environment under the condition that the recognition rate of other services is not affected, the influence of non-human voice is eliminated from the source, VAD detection is more accur