Emergency Care System using Deep Learning for People with Disabilities in Cambodia

In healthcare systems, especially those in developing countries, efficient communication between patients and medical personnel is essential. It can be very difficult for non-verbal patients to notify caregivers of emergencies, such as those who are mute or suffering from serious illnesses. In these situations, traditional communication techniques are frequently insufficient resulting in delayed reactions and compromising patient safety. In order to enable non-verbal patients to communicate their concern and request prompt assistance, this study proposes a hand-sign recognition and the position detection system. Our proposed approach improves response times and patient safety by combining deep learning with real-time video monitoring to identify particular hand gestures. Additionally, we demonstrated our custom hand sign dataset for recognizing SOS signs that indicate an emergency or a help request. According to the experience results, the solution we proposed used an LSTM-based model to recognize emergency hand signs with a 98% accuracy rate. By giving patients who require immediate care dependable nonverbal communication, our proposed solution promises to enhance critical care situations. The proposed system holds promise for improving emergency care in resource-limited hospital settings by enabling faster recognition of emergency signals. Moreover, this significant improve contributes to greater patient safety, satisfaction, and overall quality of care, especially for non-verbal individuals who traditionally face significant communication barriers in medical emergencies.