Audio Fingerprinting With Ternary Bitmaps for Fast Network Message Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio identification methods, particularly in call progress analysis, face challenges with accuracy and speed due to high computational intensity and false positives, making them unsuitable for real-time applications in outbound calls.
Innovation Solution
The development of a system that generates ternary fingerprint bitmaps for efficient comparison, using a hash key aggregation method to improve matching speed and accuracy, and integrates acoustic fingerprinting with conventional Answering Machine Detection (AMD) algorithms to identify network messages during outbound calls.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional AMD algorithms are used for audio identification, then the system can detect answering machines, but the computational intensity is high and false positives occur
Solution Approach 1:
The audio signal is divided into multiple frames, and each frame is processed independently to generate fingerprint bits. This segmentation allows the system to process audio in manageable chunks, reducing overall computational intensity while maintaining identification accuracy through cumulative frame analysis
Solution Approach 2:
The patent extracts only the essential spectral features (energy differences between frequency bins) needed for identification, discarding redundant audio data. This extraction approach reduces computational complexity by focusing only on the most discriminative features for detecting answering machines
2Measurement precision
If spectral analysis is used to distinguish speech from background noise, then Voice Activity Detection accuracy improves, but processing speed decreases
Solution Approach 1:
The patent extracts only the essential spectral features (energy differences between frequency bins) needed for identification, discarding redundant audio data. This extraction approach reduces computational complexity by focusing only on the most discriminative features for detecting answering machines
Solution Approach 2:
The system performs preliminary spectral analysis on each audio frame to extract energy differences before proceeding to fingerprint generation and comparison. This preliminary extraction of critical features enables faster subsequent processing while maintaining detection accuracy
3Productivity
If AMD algorithms are made more accurate to reduce false positives, then call throughput increases, but response time increases due to additional processing
Solution Approach 1:
The audio signal is divided into multiple frames, and each frame is processed independently to generate fingerprint bits. This segmentation allows the system to process audio in manageable chunks, reducing overall computational intensity while maintaining identification accuracy through cumulative frame analysis
Solution Approach 2:
The patent creates a compact fingerprint representation (copy) of the audio signal that captures essential identification features in a condensed form. This fingerprint copy enables rapid comparison against known answering machine patterns without requiring full spectral analysis of the entire audio stream
Data Source
AI summary
Systems and methods for the matching of datasets, such as input audio segments, with known datasets in a database are disclosed. In an illustrative embodiment, the use of the presently disclosed systems and methods is described in conjunction with recognizing known network message recordings encountered during an outbound telephone call. The methodologies include creation of a ternary fingerprint bitmap to make the comparison process more efficient. Also disclosed are automated methodologies for creating the database of known datasets from a larger collection of datasets.