Open AccessResearch on 3D Object Detection Method Based on Convolutional Attention Mechanism
Author(s) -
Yong Zhang,
Xiaoxia Zhang,
Da Nana
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1848/1/012097
Subject(s) - computer science , artificial intelligence , point cloud , feature extraction , pattern recognition (psychology) , computer vision , object detection , rgb color model , normalization (sociology) , feature (linguistics) , object (grammar) , linguistics , philosophy , sociology , anthropology
In 3D object detection, the illumination and occlusion of the input point cloud data lead to inaccurate feature extraction, and the maximum pooling method destroys the information structure of the point cloud, leading to the problem of weak local feature expression. This paper proposes a 3D object detection method based on Convolutional Attention Mechanism (CAM). CAM first adds an attention mechanism to the first and last layers of the traditional feature extraction network structure, then fuses the feature information of different layers, and finally performs normalization operations. Experimental results show that this method has achieved better results on the KITTI and SUN-RGB data sets compared with mainstream algorithms DoBEM, MV3D, DSS, COG, 2D-driven, and FPNet. The mAP index has increased by 0.6%-12.5%. While CAM realizes the fusion of local and global information, it significantly improves the accuracy of object detection in illuminated and occluded scenes.