Design of hybrid quantum convolutional neural network for ottoman character recognition

Spread over three continents, the texts inherited by the Ottoman Empire, which hosted different ethnic elements in a vast geography, are the subject of language and history research. One of the most significant difficulties encountered during Ottoman language research is reading handwritten texts. Especially during the reading of handwritten texts, researchers may encounter various problems. The shape of the letters according to their position in the sentence and the different cursive formats require deep expertise to read the texts. Various machine learning methods are available to solve this problem, such as a character recognition application. In previous studies, printed characters have been recognized as having high performance. However, no high-performance results have been obtained for recognizing handwritten characters. Therefore, in this study, we first design and compare hybrid quanvolutional neural networks that use quantum computing, a type of quantum machine learning that can provide superiority