Searching Efficient 3D Architectures with Sparse Point-Voxel Convolution

07/31/2020
by   Haotian Tang, et al.
4

Self-driving cars need to understand 3D scenes efficiently and accurately in order to drive safely. Given the limited hardware resources, existing 3D perception models are not able to recognize small instances (e.g., pedestrians, cyclists) very well due to the low-resolution voxelization and aggressive downsampling. To this end, we propose Sparse Point-Voxel Convolution (SPVConv), a lightweight 3D module that equips the vanilla Sparse Convolution with the high-resolution point-based branch. With negligible overhead, this point-based branch is able to preserve the fine details even from large outdoor scenes. To explore the spectrum of efficient 3D models, we first define a flexible architecture design space based on SPVConv, and we then present 3D Neural Architecture Search (3D-NAS) to search the optimal network architecture over this diverse design space efficiently and effectively. Experimental results validate that the resulting SPVNAS model is fast and accurate: it outperforms the state-of-the-art MinkowskiNet by 3.3 SemanticKITTI leaderboard. It also achieves 8x computation reduction and 3x measured speedup over MinkowskiNet with higher accuracy. Finally, we transfer our method to 3D object detection, and it achieves consistent improvements over the one-stage detection baseline on KITTI.

READ FULL TEXT
research
04/25/2022

PVNAS: 3D Neural Architecture Search with Point-Voxel Convolution

3D neural networks are widely used in real-world applications (e.g., AR/...
research
07/08/2019

Point-Voxel CNN for Efficient 3D Deep Learning

We present Point-Voxel CNN (PVCNN) for efficient, fast 3D deep learning....
research
04/30/2021

Multi Voxel-Point Neurons Convolution (MVPConv) for Fast and Accurate 3D Deep Learning

We present a new convolutional neural network, called Multi Voxel-Point ...
research
09/08/2021

RepNAS: Searching for Efficient Re-parameterizing Blocks

In the past years, significant improvements in the field of neural archi...
research
09/06/2021

Voxel Transformer for 3D Object Detection

We present Voxel Transformer (VoTr), a novel and effective voxel-based T...
research
07/20/2022

EASNet: Searching Elastic and Accurate Network Architecture for Stereo Matching

Recent advanced studies have spent considerable human efforts on optimiz...
research
07/23/2020

Representation Sharing for Fast Object Detector Search and Beyond

Region Proposal Network (RPN) provides strong support for handling the s...

Please sign up or login with your details

Forgot password? Click here to reset