We present SSS3D, a fast multi-objective NAS framework designed to find computationally efficient 3D semantic scene segmentation networks. It uses RandLA-Net, an off-the-shelf point-based network, as a super-network to enable weight sharing and reduce search time by 99.67% for single-stage searches. SSS3D has a complex search space composed of sampling and architectural parameters that can form 2.88 * 10^17 possible networks. To further reduce search time, SSS3D splits the complete search space and introduces a two-stage search that finds optimal subnetworks in 54% of the time required by single-stage searches.
@article{arxiv.2304.11207,
title = {SSS3D: Fast Neural Architecture Search For Efficient Three-Dimensional Semantic Segmentation},
author = {Olivier Therrien and Marihan Amein and Zhuoran Xiong and Warren J. Gross and Brett H. Meyer},
journal= {arXiv preprint arXiv:2304.11207},
year = {2023}
}
Comments
Accepted as a full paper by the TinyML Research Symposium 2023