ElHa Net: A Pattern Recognition Neural Network

Neural Networks are the state-of-the-art models that derive intelligent systems. Generalization of the neural network model is a judge of how well an architecture mimic human intelligence. Self-extraction neural networks become the dominant neural networks, especially in autonomous systems such as intelligent robotics, autonomous vehicles, ⃛ etc. This paper presents a neural network based on distance measure, ElHaNet Neural Network. ElHaNet network is a self-extraction neural network and is capable of pattern recognition. ElHaNet architecture composed of an extraction phase and a classification phase. Neurons in the extraction network capture patterns in the presented input. The classification network performs on the input patterns to be classified according to the problem in hand. The developed architecture is bench marked against well-known architectures for hand written digit recognition. The performance of the network is found to compete favorably with state-of-the-art models in the literature.