The impact of bio-inspired models on edge detection: a survey

Edges constitute a fundamental low-dimensional visual structure that plays a critical role in complex image analysis and comprehension. Their detection serves as an essential initial step in biometric object recognition and subsequent information processing. This paper critically reviews bio-inspired edge detection methods, which draw inspiration from the mechanisms of the human visual system, including classical and non-classical receptive fields (CRFs and nCRFs), hierarchical pathways, and contextual modulation. Traditional methods are lightweight and effective in noise suppression but lack scalability, global context integration, and biological complexity. Bio-inspired deep learning approaches combine these principles with neural networks, enhancing edge detection through hierarchical simulations, multi-scale processing, and feedback mechanisms. In addition, we review the most used metrics to evaluate the performance of the methods. This analysis highlights the strengths and limitations of traditional and hybrid approaches, providing insights into their potential to advance detection in complex visual environments.