Early methods for generating ceramic image descriptions had issues with insufficient accuracy in recognition and description.To address these issues,this paper proposes a multi-scale image feature extraction method based on deep residual networks and feature pyramid networks,and utilizes a long short-term memory network with additive attention mechanism to generate a Res FL model for Chinese descriptions.The experimental results show that the Res FL model is significantly superior to traditional neural network methods in terms of description accuracy and detail capture,and has high application value in improving the consistency and accuracy of ceramic image description.
维普
万方数据