Audio retrieval methods, devices, electronic devices and storage media

CN118312638BActive Publication Date: 2024-12-31TENCENT TECHNOLOGY (SHENZHEN) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410565110.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-09
Publication Date
2024-12-31
Estimated Expiration
2044-05-09

AI Technical Summary

Technical Problem

Existing audio retrieval technologies are difficult to ensure the accuracy of retrieval in multiple sound source environments, especially in complex audio scenarios such as multi-person conversations or multi-instrumental music.

Method used

By semantic splitting the retrieved audio, a sub-audio with independent semantics is generated, and the corresponding reference audio and similarity are obtained for each sub-audio, the candidate audio is determined, and the target audio is finally selected from the candidate audio to improve the search accuracy .

Benefits of technology

Improve the accuracy and efficiency of audio retrieval, and can capture the semantics of sub-audio more accurately and narrow the search range, thereby providing more accurate matching results in complex audio scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118312638B_ABST
    Figure CN118312638B_ABST
Patent Text Reader

Abstract

The embodiment of the present application discloses an audio retrieval method, device, electronic device and storage medium; the method includes obtaining audio to be retrieved; performing semantic splitting processing on the audio to be retrieved to obtain multiple sub-audios with independent semantics; for each sub-audio, obtaining multiple reference audios corresponding to the sub-audio and the first similarity between the sub-audio and each reference audio; based on the first similarity, determining at least one candidate audio that matches the sub-audio in the multiple reference audios, and determining the second similarity between the candidate audio and the audio to be retrieved; based on the second similarity, determining the target audio from the candidate audio. In the embodiment of the present application, the audio to be retrieved is split into multiple sub-audios with independent semantics, and then gradually retrieved from the reference audio based on the sub-audios. Therefore, this scheme can improve the accuracy of audio retrieval.
Need to check novelty before this filing date? Find Prior Art