HOMEY: Heuristic Object Masking with Enhanced YOLO for Property Insurance Risk Detection

Abstract

This paper introduces HOMEY (Heuristic Object Masking with Enhanced YOLO), a novel framework for automated property insurance risk detection. By integrating domain-specific masking strategies with an enhanced YOLO architecture and a custom risk-aware loss function, the model effectively detects 17 categories of property-related risks, including structural damage, maintenance neglect, and liability hazards. Experimental results demonstrate improved detection accuracy and robustness over baseline YOLO models while maintaining real-time inference capability.