摘要
针对传统三维模型分类算法时间复杂度较高、分类准确率较低等问题,提出一种基于体素模型与卷积神经网络的三维模型分类算法。将原始模型表示为八叉树结构的体素模型以优化模型的性状表达,使用设计的卷积神经网络对体素模型进行特征提取以及分类运算。实验结果表明,与其他三维模型分类算法相比,该分类算法的显存占用较小,同时具有较低的时间复杂度和较高的分类能力。
In view of the high time complexity and low accuracy of traditional 3D shape classification algorithm,this paper proposes a 3D shape classification algorithm based on voxel model and convolutional neural network.We expressed the original model as a voxel model with octree structure to optimize the trait expression of the model,and then used the designed convolutional neural network to extract the features of the voxel model,and performed classification operations.The experimental results show that compared with the 3D model classification algorithm,our classification algorithm has less vedio memory,lower time complexity and higher classification ability.
作者
刘泽鑫
万旺根
Liu Zexin;Wan Wanggen(School of Communication and Information Engineering,Shanghai University,Shanghai 200444,China;Institute of Smart City,Shanghai University,Shanghai 200444,China)
出处
《计算机应用与软件》
北大核心
2020年第1期262-266,319,共6页
Computer Applications and Software
基金
上海市科委港澳台科技合作项目(18510760300)
