多尺度特征融合下三维视觉图像场景分割算法  

3D visual image scene segmentation algorithm based on multi-scale feature fusion

作　　者：闫景富 王鹏飞 YAN Jingfu;WANG Pengfei(China University of Petroleum(Beijing)at Karamay,Karamay 834000,China)

机构地区：[1]中国石油大学(北京)克拉玛依校区,新疆克拉玛依834000

出　　处：《现代电子技术》2024年第21期46-50,共5页Modern Electronics Technique

基　　金：新疆维吾尔自治区自然科学基金资助项目(2022D01A165)。

摘　　要：为减少噪声对分割结果的影响,降低单一尺度特征对分割结果的敏感性,提升分割算法的鲁棒性与稳定性,并增强分割边界清晰度,提高分割精度,文中提出一种多尺度特征融合下三维视觉图像场景分割算法。双路径多信息域注意力模块通过结合频域通道与空间注意力机制,提取三维视觉图像的多尺度特征,降低单一尺度特征对分割结果的敏感性;在多尺度特征融合模块内添加空洞卷积层,增大多尺度特征的感受野,并融合增大感受野的多尺度特征,捕捉图像的细节信息和全局信息,减少噪声对分割结果的影响,提升分割算法的鲁棒性与稳定性;利用Softmax分类器处理融合特征,得到三维视觉图像场景分割结果;通过全连接条件随机场、后处理分割结果,优化分割边界清晰度,提高分割精度。实验结果证明:该算法可有效提取三维视觉图像的多尺度特征,有效完成三维视觉图像场景分割,且场景分割的边界非常清晰。为三维视觉图像的处理与分析提供了新的思路和方法。A 3D visual image scene segmentation algorithm based on multi-scale feature fusion is proposed to reduce the impact of noise on the segmentation results,reduce the sensitivity of single scale features to the segmentation results,improve the robustness and stability of the segmentation algorithm,enhance the segmentation boundary sharpness and improve the segmentation accuracy.Dual-path multi-information domain attention module is responsible for extracting multi-scale features of 3D visual images by combining frequency domain channel and spatial attention mechanism,so as to reduce the sensitivity of single scale features to segmentation results.A hollow convolution layer is added in the multi-scale feature fusion module to increase the receptive field of multi-scale features,and the multi-scale features of the enlarged receptive field are fused to capture the detailed information and global information of the image,reduce the impact of noise on the segmentation results,and improve the robustness and stability of the segmentation algorithm.The Softmax classifier is used to process the fused features to obtain the scene segmentation results of 3D visual images.By fully-connected conditional random fields(CRFs),the segmentation results are post-processed to optimize the segmentation boundary sharpness and improve the segmentation accuracy.Experimental results show that the proposed algorithm can effectively extract multi-scale features of 3D visual images and segment the scene of 3D visual images,and the boundary of scene segmentation is very clear.It is a new idea for the processing and analysis of 3D visual images.

关 键 词：多尺度 特征融合 三维视觉 图像场景分割 注意力机制 空洞卷积 Softmax分类器 条件随机场 

分 类 号：TN911.73-34[电子电信—通信与信息系统] TP391.41[电子电信—信息与通信工程]