Classification of cries of infants with cleft-palate using parallel hidden Markov models
Dror Lederman1, Ehud Zmora, Stephanie Hauschildt
1Department of ECE, Ben-Gurion University of the Negev, Beer-Sheva, Israel. drorle@ee.bgu.ac.il
Insights
A new algorithm using parallel hidden Markov models (PHMM) accurately classifies infant cries related to cleft palate (CP). This method significantly improves upon traditional hidden Markov models (HMM) for diagnosing CP in infants.
Area of Science:
- Medical research
- Computational linguistics
- Infant health
Background:
- Infant cry analysis is crucial for diagnosing medical conditions.
- Cleft palate (CP) in infants can affect vocalizations, necessitating accurate classification methods.
- Traditional hidden Markov models (HMM) have limitations in analyzing age-dependent vocal patterns.
Purpose of the Study:
- To develop and evaluate a novel algorithm for classifying infant cries associated with cleft palate (CP).
- To introduce a parallel hidden Markov model (PHMM) designed to overcome age-related variations in infant cries.
- To compare the performance of the proposed PHMM algorithm against standard HMM for CP infant cry classification.
Main Methods:
- Development of a hidden Markov model (HMM)-based cry classification algorithm.
- Introduction of a parallel HMM (PHMM) incorporating a maximum-likelihood decision rule to address age masking.
- Evaluation of algorithm performance using a database of cries from infants with cleft palate (CLP), testing various model parameters and feature sets.
Main Results:
- The proposed algorithm achieved an average correct classification rate of 91% in subject- and age-dependent experiments.
- The parallel hidden Markov model (PHMM) demonstrated significantly superior performance compared to the standard HMM.
- PHMM effectively improved the classification accuracy for cries from infants with cleft palate across different ages.
Conclusions:
- The developed PHMM algorithm offers a robust and accurate method for classifying infant cries related to cleft palate.
- PHMM technology provides a significant advancement over traditional HMM for analyzing age-variable vocalizations in infants with CP.
- This approach holds promise for improved diagnostic tools in infant healthcare, specifically for identifying cleft palate.
Abstract:
This paper addresses the problem of classification of infants with cleft palate. A hidden Markov model (HMM)-based cry classification algorithm is presented. A parallel HMM (PHMM) for coping with age masking, based on a maximum-likelihood decision rule, is introduced. The performance of the proposed algorithm under different model parameters and different feature sets is studied using a database of cries of infants with cleft palate (CLP). The proposed algorithm yields an average of 91% correct classification rate in a subject- and age-dependent experiment. In addition, it is shown that the PHMM significantly outperforms the HMM performance in classification of cries of CLP infants of different ages.
