Automated discomfort detection for premature infants in NICU using time-frequency feature-images and CNNs

Yue Sun; Deedee Kommers; Tao Tan; Wenjin Wang; Xi Long; Caifeng Shan; Carola van Pul; Ronald M. Aarts; Peter Andriessen; Peter H.N. de With

doi:10.1117/12.2549250

Automated discomfort detection for premature infants in NICU using time-frequency feature-images and CNNs

Yue Sun, Deedee Kommers, Tao Tan, Wenjin Wang, Xi Long, Caifeng Shan, Carola van Pul, Ronald M. Aarts, Peter Andriessen, Peter H.N. de With

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

Samenvatting

Cumulative discomfort of preterm infants can lead to abnormal development. An automated video-based discomfort detection system is proposed by analyzing motion patterns. We employ optical flow to estimate body motion across video-frames. Log Mel-spectrogram, Mel Frequency Cepstral Coefficients, and Spectral Subband Centroid Frequency features are calculated from 1D motion signals. These features enable 1D motion signals to be represented by 2D time-frequency images. Finally, deep CNNs are used on the 2D images for comfort/discomfort classification. The model was evaluated using leave-one-infant-out cross-validation on 183 video segments recorded during 17 heel prick events. Experimental results showed an AUC value of 0.985.

Originele taal-2	Engels
Titel	SPIE Medical Imaging
Uitgeverij	SPIE
Aantal pagina's	8
Volume	11314
DOI's	https://doi.org/10.1117/12.2549250
Status	Gepubliceerd - 2020
Evenement	2020 SPIE Medical Imaging: Image-Guided Procedures, Robotic Interventions, and Modeling - Houston, Verenigde Staten van Amerika Duur: 16 feb 2020 → 19 feb 2020

Congres

Congres	2020 SPIE Medical Imaging
Land	Verenigde Staten van Amerika
Stad	Houston
Periode	16/02/20 → 19/02/20

Vingerafdruk

Bibliografische nota

Conference 11314

Citeer dit

Sun, Y., Kommers, D., Tan, T., Wang, W., Long, X., Shan, C., ... de With, P. H. N. (2020). Automated discomfort detection for premature infants in NICU using time-frequency feature-images and CNNs. In SPIE Medical Imaging (Vol. 11314). [11314] SPIE. https://doi.org/10.1117/12.2549250

@inproceedings{358fa9abc61045d698d50f16b015dd14,

title = "Automated discomfort detection for premature infants in NICU using time-frequency feature-images and CNNs",

abstract = "Cumulative discomfort of preterm infants can lead to abnormal development. An automated video-based discomfort detection system is proposed by analyzing motion patterns. We employ optical flow to estimate body motion across video-frames. Log Mel-spectrogram, Mel Frequency Cepstral Coefficients, and Spectral Subband Centroid Frequency features are calculated from 1D motion signals. These features enable 1D motion signals to be represented by 2D time-frequency images. Finally, deep CNNs are used on the 2D images for comfort/discomfort classification. The model was evaluated using leave-one-infant-out cross-validation on 183 video segments recorded during 17 heel prick events. Experimental results showed an AUC value of 0.985.",

author = "Yue Sun and Deedee Kommers and Tao Tan and Wenjin Wang and Xi Long and Caifeng Shan and {van Pul}, Carola and Aarts, {Ronald M.} and Peter Andriessen and {de With}, {Peter H.N.}",

note = "Conference 11314",

year = "2020",

doi = "10.1117/12.2549250",

language = "English",

volume = "11314",

booktitle = "SPIE Medical Imaging",

publisher = "SPIE",

address = "United States",

}

Sun, Y , Kommers, D , Tan, T , Wang, W , Long, X , Shan, C , van Pul, C , Aarts, RM , Andriessen, P & de With, PHN 2020, Automated discomfort detection for premature infants in NICU using time-frequency feature-images and CNNs. in SPIE Medical Imaging. vol. 11314, 11314, SPIE, 2020 SPIE Medical Imaging, Houston, Verenigde Staten van Amerika, 16/02/20. https://doi.org/10.1117/12.2549250

Automated discomfort detection for premature infants in NICU using time-frequency feature-images and CNNs. / Sun, Yue ; Kommers, Deedee ; Tan, Tao ; Wang, Wenjin ; Long, Xi ; Shan, Caifeng ; van Pul, Carola ; Aarts, Ronald M.; Andriessen, Peter ; de With, Peter H.N.

SPIE Medical Imaging. Vol. 11314 SPIE, 2020. 11314.

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

TY - GEN

T1 - Automated discomfort detection for premature infants in NICU using time-frequency feature-images and CNNs

AU - Sun, Yue

AU - Kommers, Deedee

AU - Tan, Tao

AU - Wang, Wenjin

AU - Long, Xi

AU - Shan, Caifeng

AU - van Pul, Carola

AU - Aarts, Ronald M.

AU - Andriessen, Peter

AU - de With, Peter H.N.

N1 - Conference 11314

PY - 2020

Y1 - 2020

N2 - Cumulative discomfort of preterm infants can lead to abnormal development. An automated video-based discomfort detection system is proposed by analyzing motion patterns. We employ optical flow to estimate body motion across video-frames. Log Mel-spectrogram, Mel Frequency Cepstral Coefficients, and Spectral Subband Centroid Frequency features are calculated from 1D motion signals. These features enable 1D motion signals to be represented by 2D time-frequency images. Finally, deep CNNs are used on the 2D images for comfort/discomfort classification. The model was evaluated using leave-one-infant-out cross-validation on 183 video segments recorded during 17 heel prick events. Experimental results showed an AUC value of 0.985.

AB - Cumulative discomfort of preterm infants can lead to abnormal development. An automated video-based discomfort detection system is proposed by analyzing motion patterns. We employ optical flow to estimate body motion across video-frames. Log Mel-spectrogram, Mel Frequency Cepstral Coefficients, and Spectral Subband Centroid Frequency features are calculated from 1D motion signals. These features enable 1D motion signals to be represented by 2D time-frequency images. Finally, deep CNNs are used on the 2D images for comfort/discomfort classification. The model was evaluated using leave-one-infant-out cross-validation on 183 video segments recorded during 17 heel prick events. Experimental results showed an AUC value of 0.985.

U2 - 10.1117/12.2549250

DO - 10.1117/12.2549250

M3 - Conference contribution

VL - 11314

BT - SPIE Medical Imaging

PB - SPIE

ER -

Toegang tot document

10.1117/12.2549250

Gerelateerde inhoud

Impacts

Perinatal Medicine

Impact: Research Topic/Theme (at group level)