The demand for intelligent monitoring systems tailored to elderly living environments is rapidly increasing worldwide with population aging. Traditional acoustic scene monitoring systems that rely on cloud computing are limited by data transmission delays and privacy concerns. Hence, this study proposes an acoustic scene recognition system that integrates edge computing with deep learning to enable real-time monitoring of elderly individuals’ daily activities. The system consists of low-power edge devices equipped with multiple microphones, portable wearable components, and compact power modules, ensuring its seamless integration into the daily lives of the elderly. We developed four deep learning models—convolutional neural network, long short-term memory, bidirectional long short-term memory, and deep neural network—and used model quantization techniques to reduce the computational complexity and memory usage, thereby optimizing them to meet edge device constraints. The CNN model demonstrated superior performance compared to the other models, achieving 98.5% accuracy, an inference time of 2.4 ms, and low memory requirements (25.63 KB allocated for Flash and 5.15 KB for RAM). This architecture provides an efficient, reliable, and user-friendly solution for real-time acoustic scene monitoring in elderly care.

This study proposes an edge computing-based facial expression recognition system that is low cost, low power, and privacy preserving. It utilizes a minimally obtrusive cap-based system designed for the continuous and real-time monitoring of a user’s facial expressions. The proposed method focuses on detecting facial skin deformations accompanying changes in facial expressions. A multi-zone time-of-flight (ToF) depth sensor VL53L5CX, featuring an 8 × 8 depth image, is integrated into the front brim of the cap to measure the distance between the sensor and the user’s facial skin surface. The distance values corresponding to seven universal facial expressions (neutral, happy, disgust, anger, surprise, fear, and sad) are transmitted to a low-power STM32F476 microcontroller (MCU) as an edge device for data preprocessing and facial expression classification tasks utilizing an on-device pre-trained deep learning model. Performance evaluation of the system is conducted through experiments utilizing data collected from 20 subjects. Four deep learning algorithms, including Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) networks, and Deep Neural Networks (DNN), are assessed. These algorithms demonstrate high accuracy, with CNN yielding the best result, achieving an accuracy of 89.20% at a frame rate of 15 frames per second (fps) and a maximum latency of 2 ms.