Notes on edge detection approaches

Edge detection is an important research area that finds widespread applications in various fields, like image segmentation, shape extraction, pattern recognition, medical image processing, and motion analysis, etc. It is a mathematical model that identifies points in a digital image at which the intensities of an image changes significantly are known as edges or region boundaries. However, it is a critical concern that what is the minimum value of significant intensity change or the threshold for a situation. To cope with this concern, different edge detection methods are being developed. But still, this concern is not completely solved—as the problem is ill-posed. This article briefly describes important notes on the requirements and difficulties of various edge detection approaches classifying into five categories to find accurate edges. It critically reviews the setting of thresholds by different techniques through investigating the performances to find the state-of-the-art. In addition, it points out the current challenges and shows possible future research directions.