ADAPTIVE MULTI-KERNEL BACKGROUND SUBTRACTION USING FUZZY LOGIC FOR ROBUST OBJECT DETECTION IN COMPLEX SCENES

Abstract This research introduces an adaptive background subtraction method utilizing Fuzzy Logic. The approach dynamically adjusts to factors such as illumination variations, scene complexity, and motion dynamics, ensuring the optimal model selection. By leveraging Fuzzy rules, various background subtraction models are managed in real-time, and parameters are updated adaptively. The proposed method enhances stability and accuracy in object detection systems. Experimental results demonstrate that the Fuzzy Logic-based background subtraction technique exhibits greater flexibility and efficiency compared to conventional and deep learning-based methods.