Abstract: Deep learning models in computer vision face challenges such as high computational resource demands and limited generalization in practical scenarios. To address these issues, this study ...