Efficient Automatic Driving Instance Segmentation Method Based on Detection

被引：0
作者：
Chen Y. [1 ]
Wang H. [1 ]
Cai Y. [2 ]
Chen L. [2 ]
Li Y. [2 ]
机构：
[1] School of Automotive and Traffic Engineering, Jiangsu University, Zhenjiang
[2] Institute of Automotive Engineering, Jiangsu University, Zhenjiang
来源：
Qiche Gongcheng/Automotive Engineering | 2023年 / 45卷 / 04期
关键词：
autonomous vehicles; deep Learning; instance segmentation; object detection;
D O I：
10.19562/j.chinasae.qcgc.2023.04.002
中图分类号：
学科分类号：
摘要：
The instance segmentation algorithm based on deep learning has achieved excellent perfor⁃ mance in large-scale general scenarios. However，the segmentation of multi-objective instances for complex traffic scenes is still challenging，especially in the balance between high accuracy and fast inference speed，which is cru⁃ cial to driving safety of intelligent vehicles. In view of this，based on the real-time algorithm Orienmask，a multi-head segmentation framework is proposed based on the one-stage detection method. Specifically，the proposed framework comprises of a backbone，a feature fusion module and a multi-head mask construction module. Firstly，complete high-dimensional feature maps are obtained by adding residual structures to the backbone.Secondly，in or⁃ der to generate discriminative feature representations，the feature pyramid module is reconstructed by introducing in self-calibrate convolutions and the information propagation path is improved by global attention mechanism，so as to further optimize the feature fusion module of the proposed framework. Finally，a multi-head mask construction mech⁃ anism is proposed to significantly improve the segmentation performance of different targets by refining the size dis⁃ tribution of instances in the traffic scenes. The proposed algorithm has been tested and validated on the open-source dataset BDD100k，and has achieved an average intersection ratio of 23.3% and 19.4%（mAP@0.5：0.95）on bound⁃ ing boxes and segmentation masks，respectively. Compared with the baseline，the average index are increased by 5.2 % and 2.2 %. At the same time，the road experiment on the self-built real-vehicle platform also proves that the proposed algorithm can adapt to actual driving environments and meet the demands of real-time segmentations. © 2023 SAE-China. All rights reserved.
引用
收藏
页码：541 / 550
页数：9
相关论文
共 39 条
[31]  
LIN T Y, Et al., Feature pyramid net⁃ works for object detection［C］, Proceedings of the IEEE Confer⁃ ence on Computer Vision and Pattern Recognition, (2017)
[32]  
LIU J J, CHENG M M, Et al., Improving convolutional networks with self-calibrated convolutions［C］, Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni⁃ tion, (2020)
[33]  
LIU S, QIN H，, Et al., Path aggregation network for instance segmentation［C］, Proceedings of the IEEE Conference on Com⁃ puter Vision and Pattern Recognition, (2018)
[34]  
CAI Y, Et al., YOLOv4-5D：an effective and efficient object detector for autonomous driving［J］, IEEE Transac⁃ tions on Instrumentation and Measurement, 70, pp. 1-13, (2021)
[35]  
CHEN Y，, Et al., BDD100K：a diverse driving video database with scalable annotation tooling ［J］
[36]  
ZHAO H, Et al., Pyramid scene parsing network ［C］, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, (2017)
[37]  
JAIN A K., Data clustering：50 years beyond K-means［J］, Pat⁃ tern Recognition Letters, 31, 8, pp. 651-666, (2010)
[38]  
DENG J, Et al., ImageNet：a large-scale hierarchical image database［C］, Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition, (2009)
[39]  
GCNet：non-local networks meet squeeze-excitation networks and beyond［C］, Proceedings of the IEEE/CVF International Conference on Computer Vision Work, (2019)
1234