Audio Fingerprint Matching for Real-Time Network Message Identification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current audio identification systems, particularly in call progress analysis, face challenges with high false positive rates and computational intensity, making them inefficient and inaccurate for real-time identification of network messages during outbound calls.

Innovation Solution

The implementation of a ternary fingerprint bitmap system that masks sensitive bits to noise and artifacts, combined with an optimized exhaustive search using hash keys, enables faster and more accurate matching of acoustic fingerprints, reducing latency and computational load.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional CPA algorithms with low- and high-pass frequency discriminators are used, then call progress tones can be interpreted, but false positive rates increase and accuracy decreases

Engineering Contradiction:
Improveaccuracy of network message identificationVSAvoidfalse positive rate
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent segments the audio signal into multiple overlapping frames and creates separate fingerprint bitmaps for each frame. By comparing multiple frame-level bitmaps rather than processing the entire audio signal as one unit, the system achieves more precise localization of matches and reduces false positives while maintaining high accuracy in network message identification.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If exhaustive search of acoustic fingerprints is performed, then accurate identification can be achieved, but computational intensity and processing time increase

Engineering Contradiction:
Improveidentification accuracyVSAvoidcomputational intensity
Core Design Contradiction:
Measurement precisionVSPower

Solution Approach 1:

The patent divides the exhaustive search process into frame-level segments, where each audio frame generates a fingerprint bitmap that is independently compared against database entries. This segmentation allows the system to achieve accurate identification through comprehensive comparison while reducing overall computational intensity by processing smaller, manageable units in parallel rather than analyzing the entire audio signal simultaneously.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements a two-stage comparison process: first comparing fingerprint bitmaps at the frame level for quick matching, and only performing full audio signal analysis when frame-level matches are found. This partial action approach achieves accurate identification by focusing computational resources on promising candidates rather than performing exhaustive analysis on all possible matches, thereby reducing overall computational intensity.

Inventive Principle:
Principle #16Partial or excessive action

3Reliability

If frame-level fingerprint bitmap comparison is implemented, then false negatives are reduced, but processing complexity increases

Engineering Contradiction:
Improvereduction of false negativesVSAvoidprocessing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the audio signal into multiple overlapping frames, each generating a fingerprint bitmap. By comparing these frame-level bitmaps against database entries, the system reduces false negatives because even if one frame is misidentified, other frames provide redundant verification. The segmentation approach manages processing complexity by breaking down the complex task of entire-signal analysis into simpler, independent frame-level comparisons that can be processed efficiently.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP2832080B1Method for identifying a candidate audio segment from an outbound telephone call
Publication Date: 2017.05.03 INTERACTIVE INTELLIGENCE INC
  • EP2832080B1 patent drawingFigure 1A~1C
  • EP2832080B1 patent drawingFigure 2
  • EP2832080B1 patent drawingFigure 3A

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.