English letter recognition based on adaptive optimization spiking neural P systems

The novel dynamic guider algorithm within the adaptive optimization spiking neural P system (AOSNPS) framework is employed to create an innovative English letter recognition algorithm. This algorithm utilizes adaptive learning and diversity-based adaptation to regulate movement operators, achieving an average accuracy rate of 95.11% under low-noise conditions in classifying English letters. Leveraging the AOSNPS framework, the proposed algorithm surpasses back propagation neural networks with four gradient descent strategies across varying noise levels. The paper marks the initial application of the AOSNPS algorithm to pattern recognition, capitalizing on theoretical advancements in membrane computing and combinatorial optimization. The results underscore the potential of AOSNPS’s methodological innovations in advancing pattern recognition capabilities.