search

zhulf0804 / PointPillars

A Simple PointPillars PyTorch Implementation for 3D LiDAR(KITTI) Detection.
MIT License
454 stars 112 forks source link

Quick Shape explaination #58

Closed alre5639 closed 5 months ago

alre5639 commented 11 months ago

Hello, could you please explain the shape of the output of neck(). 384,248,216 doe this mean each voxel (248 in y and 216 in x have 384 associated features?)

Afreshbird commented 10 months ago

In paper, Figure 2. Network overview. After input feature is extracted feature by backbone, the scale of feature map is (batch_size, 6C, W/2, H/2). And the C is 64, so 64*6 = 384.