Optimisation Using Levenberg-Marquardt Algorithm of Neural Networks for Iris
Autor: | Suresh Limkar, M. M. Sardeshmukh, Asim Sayed |
---|---|
Rok vydání: | 2014 |
Předmět: | |
Zdroj: | Proceedings of the International Conference on Frontiers of Intelligent Computing: Theory and Applications (FICTA) 2013 ISBN: 9783319029306 FICTA |
Popis: | This paper explores the optimisation technique of Damped Least Square Method also known as the Levenberg-Marquardt (LM) Algorithm for Iris recognition. The motive behind it is to show that even though there are many algorithms available which act as an alternative to the LM algorithm such as the simple gradient decent and other conjugate gradient methods be it the vanilla gradient decent or the Gauss Newton iteration, the LM algorithm outperforms these optimisation techniques due to the addressing of the problem by the algorithm as the Non-linear Least Square Minimisation. The results are promising and provide an insight into Iris recognition which are distinct pattern of individuals and are unique in case of every eye. |
Databáze: | OpenAIRE |
Externí odkaz: |