CNN-LSTM for Noninvasive Neonatal Opiod Withdrawal Syndrome (NOWS) Diagnosis through Infant Cry

Lu, James

Full Metadata

Overview

Year:: 2022
Contributor:: Lu, James (creator); Silverman, Harvey (Advisor); Chason, Eric (Reader); Brown University. Engineering: Electrical Sciences and Computer Engineering (sponsor)
Genre:: theses
Subject:: Machine Learning; Deep Learning; Neonatal Opiod Withdrawal Syndrome; Signal processing
Extent:: 7, 54 p.

Files

Description

Abstract:: Deep learning shifts the way to build signal processing systems from coding or model-centric to data-centric. This paper presents a system to support data-centric deep learning for signal processing. Using new data from an ongoing medical case study, the work sets the direction for objective assessment for diagnosing neonatal opioid withdrawal syndrome (NOWS) through infant cry. Our approach to the NOWS classification decision combines two deep learning models, a long short-term memory recurrent neural network (LSTM-RNN) and a convolutional neural network (CNN), so that early decisions are not made a priori. One issue for this work was obtaining sufficient data, and we are sure we did not have enough. Also, the baseline true decisions may also be questionable. Nevertheless, with the data thus obtained, we were able to achieve nearly a 90% correct classification of verification data. Realistically, however, we are virtually certain that more data will lead to different performance levels and factual assessment of the important parameters of the input data and classifier. As clinical data becomes available over time, more work can be used with this classifier to improve its performance. In addition to the difficulty of training the Acoustic Neural Network (ANN), the work addresses issues in machine learning lifecycle such as the data pipeline, the testing benchmark, performance metrics, and deployment plan.
Notes:: Thesis (Sc. M.)--Brown University, 2022

Content

Citation

Lu, James, "CNN-LSTM for Noninvasive Neonatal Opiod Withdrawal Syndrome (NOWS) Diagnosis through Infant Cry" (2022). Electrical Sciences and Computer Engineering Theses and Dissertations. Brown Digital Repository. Brown University Library. https://repository.library.brown.edu/studio/item/bdr:qfjks855/

Relations

Collection:

Electrical Sciences and Computer Engineering Theses and Dissertations

Theses and Dissertations for the Electrical Sciences and Computer Engineering department.
...