Fire anomaly detection based on low-rank adaption fine-tuning and localization using Gradient Filtering

Yicheng Qiu, Feng Sha, Li Niu, Guangyu Zhang

Abstract: Highlights•LRA-SwinCB improves performance of fire detection compared to previous methods.•LRA-SwinCB reduced trainable params, enhances adaptability to different scenarios.•Gradient Filtering detects precise localization information of fire regions.