Scholarlink

<p>In this paper, an enhanced deep convolutional neural network (DCNN) is proposed to address the challenges of accuracy and diversity in digital art image classification. This method significantly improves the feature extraction capability and model generalization performance by introducing an attention mechanism, residual connection and transfer learning. The key improvements include optimized network architecture, use of LeakyReLU activation function and fine-tuning of pre-trained models. Experimental results show that the improved DCNN performs significantly better than traditional DCNN on multiple datasets, especially when processing digital art images with complex styles and abstract forms the classification accuracy and generalization ability are significantly improved. In addition, the model also shows superiority in indicators such as specificity and Cohen's Kappa coefficient, which further verifies the effectiveness of the combination strategy. This enhanced DCNN not only has broad application prospects in the field of digital art but also provides a valuable reference for other image classification tasks.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Improved Deep Convolutional Neural Network for Digital Art Image Classification and Identification

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月6日

查看全文
  • Huidong Zhang

摘要

In this paper, an enhanced deep convolutional neural network (DCNN) is proposed to address the challenges of accuracy and diversity in digital art image classification. This method significantly improves the feature extraction capability and model generalization performance by introducing an attention mechanism, residual connection and transfer learning. The key improvements include optimized network architecture, use of LeakyReLU activation function and fine-tuning of pre-trained models. Experimental results show that the improved DCNN performs significantly better than traditional DCNN on multiple datasets, especially when processing digital art images with complex styles and abstract forms the classification accuracy and generalization ability are significantly improved. In addition, the model also shows superiority in indicators such as specificity and Cohen's Kappa coefficient, which further verifies the effectiveness of the combination strategy. This enhanced DCNN not only has broad application prospects in the field of digital art but also provides a valuable reference for other image classification tasks.