Abstract: This study proposes an image-text multimodal classification algorithm based on a combination of convolutional neural networks (CNN) and Transformer, aiming to solve the key problems in ...