AUTHOR=Zhang Chunyu , Xu Fang , Wu Chengdong , Xu Chenglong 

TITLE=Rethinking 1D convolution for lightweight semantic segmentation

JOURNAL=Frontiers in Neurorobotics

VOLUME=Volume 17 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/neurorobotics/articles/10.3389/fnbot.2023.1119231

DOI=10.3389/fnbot.2023.1119231

ISSN=1662-5218

ABSTRACT=Lightweight semantic segmentation promotes the application of semantic segmentation in tiny devices. The existing lightweight semantic segmentation network has the problems of low precision and a large number of parameters. In response to the above problems, we designed a full 1D convolutional lightweight semantic segmentation network (LSNet). The tremendous success of this network is attributed to the following three modules: 1D multi-layer space module (1D-MS), 1D multi-layer channel module (1D-MC), and flow alignment module (FA). The 1D-MS and the 1D-MC add global feature extraction operations based on the MLP idea. This module uses 1D convolutional coding, which is more flexible than MLP. It increases the global information operation, improving features' coding ability. The FA module fuses high-level and low-level semantic information, which solves the problem of precision loss caused by the misalignment of features. We designed a 1D-mixer encoder based on the transformer structure. It performed fusion encoding of the feature space information extracted by the 1D-MS module and the channel information extracted by the 1D-MC module. 1D-mixer obtains high-quality encoded features with very few parameters, which is the key to the network's success. The attention pyramid with flow alignment module (AF-FA) uses an attention pyramid (AF) to decode features and adds a FA module to solve the problem of feature misalignment. Our network requires no pre-training and only needs a 1080Ti GPU for training. It achieved 72.6mIOU and 95.6FPS on the Cityscapes dataset and 70.5mIOU and 122FPS on the Camvid dataset. We ported the network trained on the ADE2K dataset to mobile devices, and the latency of 224 ms proves the application value of the network on mobile devices. The results on the three datasets prove that the network generalization ability we designed is powerful. Our designed network achieves the best balance between segmentation accuracy and parameters compared to state-of-the-art lightweight semantic segmentation algorithms. The parameters of LSNet are only 0.62M, which is currently the network with the highest segmentation accuracy within 1M parameters.