Abstract—Object shape detection and localization techniques that utilize snake deformable models are one of the most promising image detection techniques. The binary edge maps, derived from the original image, are basically the class acted upon by the snake to extract the desired features. As a result, high and low energy content pixels are obtained. The high energy pixels are the pixels that reflect the object borders of a given image. This paper addresses a new external force that is calculated from the energy diffusion of high content pixels and is used to balance the internal forces of the snake. The proposed scheme showed better results in terms of computation speed and capture range than standard snake models. On the basis of the concavity convergence, analogous results are achieved in the proposed scheme compared with standard models.
Index Terms—Algorithms, computer vision, deformable models, feature extraction, image edge detection.
M. A. Nisirat is with the Faculty of Engineering Technology, Albalqa Applied University, Amman, Jordan (e-mail: email@example.com, firstname.lastname@example.org).
Cite: Mahdi A. Nisirat, "A New External Force for Snake Algorithm Based on Energy Diffusion," International Journal of Machine Learning and Computing vol. 9, no. 3, pp. 316-321, 2019.