基于迭代神经网络的图像结构表示和分类

由于缺少结构化的表示,基于内容的图像分类存在一定的问题,据此提出了一种基于迭 代神经网络的自然图像表示和分类的方法。利用Berkeley分割算法将图像分割成不同的区域,采用 基于人工的多叉树或基于邻接区域的二叉树的方法进行区域合并,同时提取区域统计特征,得到图像 的树型结构表示。根据BPTS算法对网络进行训练,训练好的网络就具备了图像分类的功能。实验 结果表明,基于迭代神经网络的结构表示和分类方法具有很强的结构学习能力,同时人工生成的多叉 树涵盖更多的语义信息且能得到较好的分类结果。

Image structure representation and classification based on recursive neural network

Due to the lack of structural information representation,there still exist key issues in the research domain of content-based image retrieval. In the paper, a recursive neural network-based approach for structure representation and pattern classification was presented. The natural scene images were segmented using Berkeley algorithm and integrated by multi-branch directed trees manually or binary trees automatically. The statistical features were also extracted and got the structure representation. After the training by BPTS algorithm,the structural networks can be employed into image classification. Experimental results show that the recursive network-based structure model has significant ability to represent complicated patterns and the multi-branch tree has more semantic meanings to achieve higher classification rates.