Authored by Surjeet Dalal et al., supported by Princess Nourah bint Abdulrahman University | June 2024

This peer-reviewed study introduces a hybrid deep learning approach combining transfer learning and quantization to enhance the YOLO (You Only Look Once) object detection model for smart home surveillance. Optimized for edge devices, the model achieved 98.87% accuracy across RoboFlow datasets. The paper underscores practical applications in intrusion detection, energy management, and automation, while addressing privacy and scalability. It demonstrates how efficient AI models can bolster smart home security within broader smart city frameworks.

Keywords: