Quantum Particle Swarm Optimization Based Convolutional Neural Network for Handwritten Script Recognition

Even though several advances have been made in recent years, handwritten script recognition is still a challenging task the pattern domain. This field has gained much interest lately due to its diverse application potentials. Nowadays, different methods are available for automatic recognition. Among most of reported techniques, deep neural networks achieved impressive results and outperformed classical machine learning algorithms. However, process designing such right from scratch intuitively appears incur significant amount trial error, which renders them unfeasible. approach often requires manual intervention with domain expertise consumes substantial time computational resources. To alleviate this shortcoming, paper proposes new architecture search based on meta-heuristic quantum particle swarm optimization (QPSO), capable automatically evolving meaningful convolutional network (CNN) topologies. The experiments conducted eight datasets belonging three popular Indic scripts, namely Bangla, Devanagari, Dogri, consisting characters digits. Empirically, imply that proposed QPSO-CNN algorithm outperforms state-of-the-art faster prediction higher accuracy.