MATLAB Implementation of SUSAN Corner Detection Algorithm

This is a MATLAB-implemented SUSAN corner detection program that enables users to accurately detect corner positions in digital images. The program operates by analyzing grayscale variations and gradient information within the image using a circular mask (typically 37 pixels) to identify distinctive corner features. Key implementation aspects include: the SUSAN principle that compares pixel intensities within the mask against a central nucleus, adaptive thresholding for noise robustness, and geometric moment calculations for corner response. Users can modify parameters such as the brightness threshold (usually set between 20-50) and minimum nucleus size to optimize performance for specific applications. The algorithm efficiently processes images by employing intensity comparison rather than derivative calculations, making it computationally efficient. This implementation provides foundational understanding of corner characteristics in images, serving as an accurate basis for subsequent image processing and computer vision tasks like feature matching, object recognition, and 3D reconstruction.