Infant Cry Audio Classification Using MD-GMM and Logistic Regression

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional methods for assessing infant crying, particularly distinguishing between colic and fussiness, are subjective and imprecise, relying on parental reporting which is unreliable and prone to variability.

Innovation Solution

A system that uses an audio capture device worn by the infant to record and process audio data, segmenting it into cry-related and non-cry segments, and applying a Minimum Duration Gaussian Mixture Model (MD-GMM) and binary logistic regression to classify periods as either cry or fussiness, providing objective quantification.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If parental reporting is used to assess infant crying, then the assessment can be obtained easily, but the precision and reliability of the assessment deteriorates

Engineering Contradiction:

Improveease of assessmentVSAvoidprecision of cry assessment

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The patent replaces the manual parental reporting system with an automated audio analysis system that uses signal processing and machine learning algorithms to objectively detect and classify infant crying, thereby substituting human subjectivity with computational objectivity while maintaining ease of use through automatic processing

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces an intermediary audio analysis system that acts as a mediator between the infant's crying and the assessment process, using trained classifiers and feature extraction to translate raw audio signals into reliable diagnostic information, thereby eliminating the need for direct parental interpretation

Inventive Principle:

Principle #24Intermediary (Mediator)

2Device complexity

If subjective parental assessment is used, then the device complexity is reduced, but the reliability of the assessment deteriorates

Engineering Contradiction:

Improvecomplexity of assessment systemVSAvoidreliability of cry assessment

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The patent replaces simple parental observation with an automated system using Gaussian Mixture Models and logistic regression classifiers that process audio features automatically, substituting human judgment with algorithmic decision-making that enhances reliability while managing complexity through standardized processing pipelines

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If automated audio analysis is implemented, then the measurement precision of cry assessment is improved, but the device complexity increases

Engineering Contradiction:

Improveprecision of cry quantificationVSAvoidcomplexity of audio processing system

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the audio analysis process into distinct modular components: feature extraction (MFCCs, spectral features), classification (GMM, logistic regression), and post-processing, allowing each component to be optimized independently and facilitating systematic implementation of complex audio analysis

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent transforms raw audio signals into a standardized set of acoustic parameters and features (spectral centroid, MFCCs, energy features) that can be processed by machine learning algorithms, changing the parameter space from raw waveforms to meaningful acoustic descriptors that enhance measurement precision

Inventive Principle:

Principle #35Parameter changes

4Loss of information

If conventional parental reporting is used, then the loss of information is minimized, but the measurement precision deteriorates

Engineering Contradiction:

Improveinformation retention in assessmentVSAvoidprecision of cry classification

Core Design Contradiction:

Loss of informationVSMeasurement precision

Solution Approach 1:

The patent replaces imprecise parental reporting with automated audio analysis that captures and processes the full acoustic information content of infant vocalizations, substituting human memory and interpretation limitations with computational analysis that preserves and utilizes complete acoustic data for precise classification

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11328738B2Systems and methods for automatic determination of infant cry and discrimination of cry from fussiness

Publication Date: 2022.05.10 LENA FOUNDATION

AI summary

A method including receiving one or more datasets of audio data of a key child captured in a natural sound environment of the key child. The method also includes segmenting each of the one or more datasets of audio data to create audio segments. The audio segments include cry-related segments and non-cry segments. The method additionally includes determining periods of the cry-related segments that satisfy one or more threshold non-sparsity criteria. The method further includes performing a classification on the periods to classify each of the periods as either a cry period or a fussiness period. Other embodiments are described.