The Development of Neural Network Architectures for Image Classification to Prevent Sudden Infant Death in Smart Buildings

Author

RASHIK SHADMAN, Southern Illinois University CarbondaleFollow

Date of Award

9-1-2021

Degree Name

Master of Science

Department

Electrical and Computer Engineering

First Advisor

Lu, Chao

Abstract

Sudden Infant Death Syndrome (SIDS), which means the sudden death of a baby under one year-old, is a major problem in the United States and the whole world. Normally SIDS occurs during sleep time. Among many reasons behind sudden infant death, the sleeping position is a vital factor in SIDS. Babies face breathing difficulty when they sleep on their stomach or side. It is essential to let the babies sleep on their backs, but it is not always possible. Therefore, a detection system is needed to detect the baby’s sleep position. There are various ways to do it. In our work, we used images of sleeping babies from monitoring cameras. Then we used Convolutional Neural Network architectures for the classification of those images based on the sleeping position of the babies. The highest classification accuracy was 92.5% which is comparable with the state-of-the art methods. Our proposed system is low cost and convenient. It also provides high classification accuracy. Besides, while maintaining good classification accuracy, we tried to reduce the size of the proposed CNN architectures to make our CNN architectures compatible with portable electronics.