权值优化集成卷积神经网络及其在三维模型识别中的应用

Weighted Optimization Integrated Convolutional Neural Network and Its Application in 3D Model Recognition

三维模型应用广泛,如何有效地管理和分类这些数据库中的三维模型一直是人们关注的问题。然而,由于不同三维模型之间的相似性难以测量,因而很难获得一种稳健且广泛适用的三维模型分类算法。为此,提出了一种权值优化集成卷积神经网络(WOTCNN)模型,并将其应用到三维模型的分类识别中。首先,获取三维模型的深度投影视图来最大限度地保留三维模型的空间信息。然后,采用调整的VGG网络对各角度的深度投影图像进行训练并提取预测概率值。最后,通过加权集成算法获得完整三维模型的最终分类结果。对ModelNet10及ModelNet40数据库的实验表明:三维模型的平均分类准确率达到92.84%和86.51%。在预测性能方面,该网络优于普通的单卷积神经网络;在三维模型识别方面,其分类准确率能够得到显著提升。

3 D models enjoy a popularity. It has always been our concern as to how to effectively manage and classify the 3 D models in these databases. However, due to the similarity between different 3 D models is difficult to calculate, it is difficult to obtain a robust and widely applicable 3 D model classification algorithm. Thus a weighted optimization integrated convolutional neural network model is proposed and applied to the classification and recognition of 3 D models. Firstly, the depth projection view of the 3 D model is obtained to maximize the reserve of spatial information of the 3 D model. Then, the adjusted VGG network is used to train the depth projection images from different angles and extract the predictive probability values. Finally, the final classification results of the complete 3 D model are obtained by weighted ensemble algorithm. The experiments on ModelNet10 and ModelNet40 databases show that the average classification accuracy of the 3 D model is 92.84% and 86.51% respectively. In terms of performance prediction, the network is superior to the ordinary single convolution neural network, and its classification accuracy can be significantly improved in 3 D model recognition.