Explainable AI and YOLOv8-based Framework for Indoor Fire and Smoke Detection

In recent years, the importance of Explainable AI (XAI) and computer vision has surged in vision-based problem-solving, notably in fire and smoke detection. Several models and techniques have been put forth for the detection of fire and smoke, however, they have certain shortcomings, such as high false alarm rates and poor accuracy. Our research presents a model for indoor fire and smoke detection that is based on YOLOv8. Additionally, the paper uses LIME and SHAP to show how the model learns from the images. The results from the YOLOv8n model are encouraging, with a mAP50 of 0.556 and a mAP50-95 of 0.315.