Recognition of Greek Alphabet Characters with Memristive Neuromorphic Circuit

Pattern recognition is a critical aspect of modern computing and artificial intelligence, enabling robotic and computing systems to classify and categorize information based on its patterns and relationships. In this context, memristors are emerging as a promising new type of device for pattern recognition due to their memory capabilities as non-volatile two-terminal elements and in-memory computing potential. In this work, we present a novel approach to character recognition using CBRAM memristors by combining traditional machine learning techniques with a memristor-based oscillator exhibiting neuromorphic properties. The proposed approach demonstrates the feasibility of a pattern recognition specific hardware based on memristors that can be integrated in edge computing applications, aiming to bring computation and data storage closer to the source of data, rather than relying on centralized data centers, and consequently, leading to improved latency, increased reliability, and reduced costs.