基于多模态知识对齐的双分支点云语义分割算法
|
|
杨军,卯恒睿,党吉圣
|
Dual-branch point cloud semantic segmentation algorithm based on multimodal knowledge alignment
|
|
Jun YANG,Hengrui MAO,Jisheng DANG
|
|
| 表 2 NuScenes数据集上不同算法的语义分割性能对比 |
| Tab.2 Comparison of semantic segmentation performance across different methods on NuScenes dataset |
|
| 模型 | mIoU/% | IoU/% | | 护栏 | 自行车 | 公交车 | 汽车 | 施工区域 | 摩托车 | 行人 | 交通锥 | | Cylinder3D[18] | 77.2 | 82.8 | 29.8 | 84.3 | 89.4 | 63.0 | 79.3 | 77.2 | 73.4 | | PMF[25] | 77.0 | 82.0 | 40.0 | 81.0 | 88.0 | 64.0 | 79.0 | 80.0 | 76.0 | | RViT-FLARES[22] | 77.0 | 76.7 | 39.2 | 93.0 | 92.0 | 55.2 | 81.6 | 77.2 | 64.9 | | PC-BEV[23] | 78.8 | 78.2 | 46.3 | 92.5 | 93.4 | 55.0 | 87.1 | 81.0 | 65.4 | | 2DPASS[24] | 78.5 | 76.1 | 52.6 | 95.7 | 85.1 | 56.5 | 91.7 | 80.9 | 68.2 | | 本文模型 | 79.3 | 76.5 | 52.3 | 96.1 | 88.2 | 56.6 | 91.5 | 80.6 | 69.2 | | | 模型 | IoU/% | | 拖车 | 卡车 | 可行驶区域 | 其他地面 | 人行道 | 地形 | 人造结构 | 植被 | | Cylinder3D[18] | 84.6 | 69.1 | 97.7 | 70.2 | 80.3 | 75.5 | 90.4 | 87.6 | | PMF[25] | 81.0 | 67.0 | 97.0 | 68.0 | 78.0 | 74.0 | 90.0 | 88.0 | | RViT-FLARES[22] | 70.9 | 84.1 | 96.8 | 74.1 | 95.6 | 75.1 | 88.6 | 86.7 | | PC-BEV[23] | 69.2 | 85.7 | 97.1 | 76.8 | 77.0 | 76.3 | 90.6 | 88.5 | | 2DPASS[24] | 74.6 | 82.2 | 96.6 | 76.4 | 72.4 | 71.5 | 88.2 | 87.3 | | 本文模型 | 76.5 | 88.6 | 96.5 | 75.2 | 73.2 | 72.3 | 88.0 | 87.5 |
|
|
|