Year: 2024
Type: article
Abstract: This research is in response to the rising cases of Forgotten Baby Syndrome (FBS), which has sadly led to the loss of young lives. In Malaysia, 19 incidents of child neglect in vehicles have been reported. Parents or guardians responsible for leaving a child in danger can be charged under Section 31(1)(a) of the Children Act 2001 for neglect, abuse, or abandonment. Overheating in vehicles, where the car’s temperature becomes dangerously high, is a common cause of these deaths. A system to detect children left in vehicles is important to save lives. The project starts with a detailed analysis to figure out what is needed for a strong detection system. Next, the system design phase will create a camera-based system that can quickly identify unattended children in vehicles. The development phase will use the YOLOv8s deep learning model to build and fine-tune an algorithm that can accurately detect children. Cameras will be placed in the front and back of the vehicle for full coverage. The system has been tested for its ability to detect unattended children and reduce false alarms. When using an imbalanced dataset, accuracy was 16.35%. With a balanced dataset, accuracy improved significantly to 64.48%. This ensures reliable detection in different situations, sending real-time alerts to guardians or activating the car alarm if needed. It aims to quickly notify guardians or trigger the car alarm to prevent more tragic incidents due to FBS. It provides peace of mind to parents and caregivers by protecting vulnerable children. Future improvements should involve testing in real-world situations, adding user faces to the dataset for better detection, improving its ability to handle lighting and obstacles inside the vehicle, and integrating it with the car’s power system. Adding temperature and motion sensors could also enhance the system.
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4408481911