Efficient Pattern Recognition Resource Utilization Neural Network

Neural Networks derives intelligent systems and modern autonomous applications. With the complexity introduced in today’s systems, designing architectures remains open research problem in all fields. Despite many efforts to generalize, the resources required by neural architectures remain challenge. Robots design, Smart cities, IoTs …etc. become the leading industry and driving the fourth industrial revolutions. All these challenges dictate the search for efficient utilization of resources. This paper demonstrates details of the architecture ElHa Net. The architecture finds the minimum resources required for pattern recognition domains. Composed of two stages, Extraction stage followed by classification stage. The extraction stage is a self-extraction, performed by convolutions like processes. The classification stage receives the extracted patterns and associates them through weighting to defined classes. The architecture is found to compete with many reported in literature.