Neural Transfer Learning for Cry-based Diagnosis of Perinatal Asphyxia

@article{Onu2019NeuralTL,
  title={Neural Transfer Learning for Cry-based Diagnosis of Perinatal Asphyxia},
  author={Charles C. Onu and Jonathan Lebensold and William L. Hamilton and Doina Precup},
  journal={ArXiv},
  year={2019},
  volume={abs/1906.10199}
}
Despite continuing medical advances, the rate of newborn morbidity and mortality globally remains high, with over 6 million casualties every year. The prediction of pathologies affecting newborns based on their cry is thus of significant clinical interest, as it would facilitate the development of accessible, low-cost diagnostic tools\cut{ based on wearables and smartphones}. However, the inadequacy of clinically annotated datasets of infant cries limits progress on this task. This study… 

Figures and Tables from this paper

A Multistage Heterogeneous Stacking Ensemble Model for Augmented Infant Cry Classification

TLDR
The Mel-frequency cepstral coefficients algorithm is used to generate the spectrograms and analyzed the varying feature vectors and the proposed multistage heterogeneous stacking ensemble model had the edge over the other neural network models, especially in terms of overall performance and computing power.

Who Cried When: Infant Cry Diarization with Dilated Fully-Convolutional Neural Networks

TLDR
The proposed algorithm consists of a fully-convolutional neural network (Conv-DetNet) that processes simultaneously all the audio signals acquired from the microphone in each crib and detects if the infants cried or not.

Artificial Intelligence in NICU and PICU: A Need for Ecological Validity, Accountability, and Human Factors

TLDR
The application of AI in NICU/PICU is summarized and consecutively identifies the existing flaws in AI and proposes related recommendations, which, if addressed, can improve AIs’ readiness for a real clinical environment.

Speech Emotion Recognition among Elderly Individuals using Multimodal Fusion and Transfer Learning

TLDR
This work trains machine learning models to recognize the emotions of elderly individuals via performing a 3-class classification of valence and arousal as part of the INTERSPEECH 2020 Computational Paralinguistics Challenge (COMPARE).

Towards Real-Time Multimodal Emotion Recognition among Couples

TLDR
This work seeks to develop a smartwatch-based system that leverages multimodal sensor data to recognize each partner's emotions in daily life and enable interventions to improve the emotional well-being, relationship quality, and chronic disease management of couples.

Artificial Intelligence for Healthcare in Africa

Digital technology will play a significant role in achieving sustainable human development worldwide. In 2015, United Nations Member States set 17 goals, the Sustainable Development Goals (SDGs), to

Actor Critic with Differentially Private Critic

TLDR
This work proposes a technique to achieve knowledge transfer in cases where agent trajectories contain sensitive or private information, such as in the healthcare domain, and empirically shows this technique increases sample efficiency in resource-constrained control problems while preserving the privacy of trajectories collected in an upstream task.

State and Future Prospects of Artificial Intelligence (AI) in Ghana

TLDR
The paper highlights the existing institutions leveraging AI technologies, points out some current challenges with regards to AI adoption, and identifies some exciting prospects of AI given the current state of the country.

Deep Graph Embeddings in Recommender Systems

TLDR
This thesis aims to explore the application of deep graph embeddings for recommendation, and indicates that it could be useful to bias recommender models towards long-tail items.

References

SHOWING 1-10 OF 27 REFERENCES

Harnessing infant cry for swift, cost-effective diagnosis of Perinatal Asphyxia in low-resource settings

  • Charles C. Onu
  • Medicine
    2014 IEEE Canada International Humanitarian Technology Conference - (IHTC)
  • 2014
TLDR
A support vector machine-based pattern recognition system that models patterns in the cries of known asphyxiating infants (and normal infants) and then uses the developed model for classification of `new' infants as having asphyxia or not is designed.

A System for the Processing of Infant Cry to Recognize Pathologies in Recently Born Babies with Neural Networks

TLDR
The design and implementation of the complete system that classifies three different kinds of cries from normal, deaf and asphyxiating infants, of ages from one day up to nine months old are presented.

PAIN CRY IN FULL‐TERM ASPHYXIATED NEWBORN INFANTS CORRELATED WITH LATE FINDINGS

TLDR
The results showed significant differences between the cries of the asphyxiated newborn and the healthy infants, which were more abnormal if the infant was found to be neurologically damaged at the check‐up at 2–8 years.

Assessment of infant cry: acoustic cry analysis and parental perception.

TLDR
Assessment of infant cry provides a window into the neurological and medical status of the infant and investigates the relationship between acoustic characteristics of the cry and diagnoses related to neurological damage, SIDS, prematurity, medical conditions, and substance exposure during pregnancy.

Acoustical Analysis of Pain Cries' in Neonates: Fundamental Frequency

crying comprises a rhythmic alternation of cry sounds (utterances) and inspiration. A method for knowing the reason of baby's cry is proposed using frequency analysis of the recorded voice. The

Sound Spectrographic Cry Analysis of Infants with Bacterial Meningitis

TLDR
A total of 110 pain‐induced cries from 14 infants with bacterial meningitis were analysed by sound spectrographic methods, and 15 cry attributes were studied.

Cry Characteristics of 172 Healthy 1- to 7-Day-Old Infants

TLDR
There were no significant differences noted in the cry characteristics according to gender, the age of the infants or the gestational age when they were born.

Large-Scale Video Classification with Convolutional Neural Networks

TLDR
This work studies multiple approaches for extending the connectivity of a CNN in time domain to take advantage of local spatio-temporal information and suggests a multiresolution, foveated architecture as a promising way of speeding up the training.

Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification

TLDR
This work proposes a Parametric Rectified Linear Unit (PReLU) that generalizes the traditional rectified unit and derives a robust initialization method that particularly considers the rectifier nonlinearities.

Biobehavioral Assessment of the Infant

This book is most probably already on the office shelves of infancy and early childhood researchers. The editors have done a superb job ensuring that a uniform model of reporting the varied areas