Audio Recognition Fingerprint Index Database Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional audio retrieval systems face inefficiencies due to the presence of invalid audio fingerprints resulting from audio attacks, which waste storage space and increase calculation complexity, leading to low recognition efficiency.
Innovation Solution
An audio recognition method that utilizes a classifier to remove invalid audio fingerprints from the fingerprint index database by training on feature point data before and after audio attacks, using algorithms like nearest neighbor, support vector machine, or neural networks to filter out counter-example data as invalid fingerprints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If audio fingerprints are stored in the fingerprint index database without filtering, then the database maintains completeness of audio data, but storage space is wasted and recognition efficiency decreases due to invalid fingerprints
Solution Approach 1:
The patent applies preliminary action by training a classifier in advance to identify valid versus invalid audio fingerprints before they are stored in the fingerprint index database. This pre-trained classifier then filters out invalid fingerprints during the indexing process, preventing storage waste and recognition inefficiency before they occur. The classifier is trained using audio data that has undergone various attacks (segmentation, format conversion, sampling rate conversion, dubbing) to learn patterns of valid fingerprints that withstand these transformations.
2Productivity
If audio fingerprints are filtered using a classifier, then storage space is reduced and recognition efficiency is improved, but the system complexity increases due to classifier training and maintenance
Solution Approach 1:
The classifier is trained in advance during an offline phase using comprehensive audio data that has undergone various attacks. This preliminary training allows the system to develop robust recognition capabilities without adding complexity to the real-time audio recognition process. The trained classifier is then deployed as a fixed component that simply evaluates fingerprints during normal operation, minimizing the complexity burden on the runtime system.
Data Source
AI summary
An audio recognition method, including: acquiring an audio file to be recognized (S100); extracting audio feature information of the audio file to be recognized, the audio feature information including audio fingerprints (S200); searching, in a fingerprint index database, audio attribute information matched with the audio feature information, the fingerprint index database including an audio fingerprint set in which invalid audio fingerprint removal has been performed on audio sample data (S300). As the audio fingerprint set in the fingerprint index database has been subjected to invalid audio fingerprint removal of audio sample data, the storage space of audio fingerprints in the fingerprint index database can be reduced, and the audio recognition efficiency can be improved. Further provided are an audio recognition device and a server.