Orthogonal
Polynomials
Polynomials
Descripción de la línea de investigación
2015 Gegenbauer Orthogonal Polynomias for automated image recognition / Polinomios Ortogonales de Gegenbauer para el reconocimiento automatizado de imágenes ▹ (PR) (OP) (AV)
MCC2015B -HAJA: Automated image recognition is a key topic in computer science and has been addressed with many techniques, often with good results. Yet, recent methods can still be improved: many do not reach their full potential, and alternative tools that might yield better performance remain under‑used. We therefore propose a hybrid approach that combines two state‑of‑the‑art techniques for automatic image recognition. First, the SURF algorithm is employed to extract regions of interest; second, Gegenbauer orthogonal polynomials are used to build robust, low‑redundancy image descriptors.
Experiments were carried out on 400 grayscale images depicting 200 distinct scenes. The task was to recognize each scene despite variations in scale, translation, and rotation by relying on the detected regions of interest. Results obtained with the proposed methodology were compared with those from the plain SURF algorithm using the Wilcoxon signed‑rank test. The hybrid method shows superior performance and constitutes a simple, well‑grounded framework that opens avenues for future research.
MCC2015B-HAJA: Polinomios Ortogonales de Gegenbauer para el reconocimiento automatizado de imágenes. ▹ (PR) (OP) (AV)