作者：
Cheng, JunZhang, ShengChinese Acad Sci
Shenzhen Inst Adv Technol CAS Key Lab Human Machine Intelligence Synergy Sy Shenzhen Peoples R China Wuhan Univ
Sch Elect & Automat Dept Artificial Intelligence & Automat Wuhan Peoples R China
Among various indoor3dobjectdetection methods, one of which is based on 3d convolution. However, in current 3d convolutional method for indoor3dobjectdetection, feature fusion is only implemented in a single dir...
详细信息
Among various indoor3dobjectdetection methods, one of which is based on 3d convolution. However, in current 3d convolutional method for indoor3dobjectdetection, feature fusion is only implemented in a single direction and adjacent layer, which leads to the feature loss during feature propagation and weakens the detection performance. To solve the above problem, a bidirectionalmultilayerfusionnetwork (BMFN) is proposed to enrich features for prediction. Furthermore, we propose a new network based on BMFN for 3dobjectdetection, BMFN3d. In BMFN3d, the channel and spatial attention mechanisms are introduced into the backbone and head of detectionnetwork, so as to enhance the selection of effective feature information. Finally, a new calculation method for regression loss is proposed with the consideration of the 3d spatial information. We validate our method with two indoordatasets of SUN RGB-d and ScanNet, and achieve state-of-the-art results.
暂无评论