This research presents an approach to design deep quantum neural networks with data re-uploading, quantum feedforward and quantum convolution methods. The main goal of this research is first the architecture and then the training and eva‎luation of the performance of these networks in a three-dimensional shape classification problem. Here, the training process involves optimizing the parameters of quantum neural networks to achieve optimal performance. To eva‎luate the effectiveness of the proposed quantum methods, a comparison is made with two classical neural networks that are commonly used in similar problems. The results show that quantum neural networks have achieved accuracy comparable to classical methods and it can be hoped that they will be able to improve classification accuracy in processing more complex data. This research contributes to the development of knowledge in the field of applications of quantum computing in machine learning and a better understanding of the capabilities of quantum neural networks in the classification of three-dimensional shapes.

فرهاد فضيله

وحيد سالاري

كيوان آقابابائي ساماني , حامد بخشيان