发帖
雷达卡
Point-Voxel CNN for Efficient 3D Deep Learning
Zhijian Liu Haotian Tang Yujun Lin Song Han
MIT Shanghai Jiao Tong University MIT MIT
Abstract
We present Point-Voxel CNN (PVCNN) for efficient, fast 3D deep learning. Previ-
ous work processes 3D data using either voxel-based or point-based NN models.
However, both approaches are computationally inefficient. The computation cost
and memory foot ...
京公网安备 11010802022788号
