Silent Distress Detection for Hospital Using Unintentional Human Micro-Actions

Abstract

Silent distress among hospitalized patients, particularly in intensive care units (ICUs), often goes unnoticed due to the inability of patients to communicate discomfort effectively. Traditional monitoring systems rely mainly on physiological parameters and manual observation, which may fail to capture subtle behavioral cues. This paper proposes an intelligent and non-invasive system for detecting silent distress using unintentional human microactions such as facial expressions, eye movements, posture changes, and involuntary body motions. The system utilizes computer vision and deep learning techniques, including Convolutional Neural Networks (CNN) for spatial feature extraction and Long Short-Term Memory (LSTM) models for temporal analysis. The system processes real-time video data to classify patient conditions into normal, mild distress, and critical distress levels. An automated alert mechanism notifies healthcare staff when abnormal patterns are detected. The proposed system improves patient safety, reduces response time, and enhances monitoring efficiency in hospital environments.