A Real-Time Road Scene Semantic Segmentation Model Based on Spatial Context Learning

被引:0
作者
Xiao, Xiaomei [1 ]
Tang, Jialiang [1 ]
Lu, Xiaoyan [1 ]
Feng, Zhengyong [1 ]
Li, Yi [2 ]
机构
[1] China West Normal Univ, Elect Informat Proc Engn Technol Res Ctr, Sch Elect Informat Engn, Nanchong 637009, Peoples R China
[2] Chengdu Normal Univ, Coll Phys & Engn Technol, Chengdu 611130, Peoples R China
来源
IEEE ACCESS | 2024年 / 12卷
关键词
Feature extraction; Semantics; Semantic segmentation; Accuracy; Computational modeling; Real-time systems; Training; Context modeling; Attention mechanisms; Encoding; Real-time semantic segmentation; spatial context guidance; feature attention; feature alignment; NETWORK;
D O I
10.1109/ACCESS.2024.3503676
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
To address the issues of high computational complexity and insufficient aggregation of global and local information in existing image segmentation methods, this paper proposes an efficient segmentation model based on Spatial Context Learning, named SCLSeg. The main idea is to aggregate local regions into higher-level semantic regions in a learnable manner. The proposed Spatial Context Guided Feature Alignment module (SC-FA) learns aligned features from image-level to local regions, exploring and integrating contextual information. During training, a multi-scale strategy is used to group semantic regions, and a Channel Aggregation Block (CAB) is designed to dynamically capture semantic groups through a mechanism of feature separation and fusion, thereby aggregating multi-level pixel features to generate the final segmentation results. We further introduce a boundary loss to optimize the accuracy of segmentation edges. To meet real-time processing requirements, a series of lightweight strategies and simplified structures are adopted to reduce computational costs, including lightweight encoding, channel compression, and simplified neck. Our method achieves good performance on the Cityscapes and Camvid datasets, specifically achieving 76.45% mIoU & 237 FPS on the Cityscapes test set, and 73.95% mIoU & 300.4 FPS on the CamVid test set.
引用
收藏
页码:178495 / 178506
页数:12
相关论文
共 50 条
[41]   Recognition of unstructured field road scene based on semantic segmentation model [J].
Meng Q. ;
Yang X. ;
Zhang M. ;
Guan H. .
Nongye Gongcheng Xuebao/Transactions of the Chinese Society of Agricultural Engineering, 2021, 37 (22) :152-160
[42]   Semantic Segmentation for Road Scene Based on Multiscale Feature Fusion [J].
Yi Qingming ;
Zhang Wenting ;
Shi Min ;
Shen Jialin ;
Luo Aiwen .
LASER & OPTOELECTRONICS PROGRESS, 2023, 60 (12)
[43]   HoloParser: Holistic Visual Parsing for Real-Time Semantic Segmentation in Autonomous Driving [J].
Li, Shu ;
Yan, Qingqing ;
Shi, Wenbo ;
Wang, Liuyi ;
Liu, Chengju ;
Chen, Qijun .
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2023, 72
[44]   Multi Path Real-time Semantic Segmentation Network in Road Scenarios [J].
Pengfei, Gao ;
Xiaolong, Tian ;
Cuihong, Liu ;
Chenfei, Yang .
INTERNATIONAL JOURNAL OF INTELLIGENT TRANSPORTATION SYSTEMS RESEARCH, 2024, :47-60
[45]   Real-Time Semantic Segmentation A brief survey and comparative study in remote sensing [J].
Broni-Bediako, Clifford ;
Xia, Junshi ;
Yokoya, Naoto .
IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE, 2023, 11 (04) :94-124
[46]   Limited Receptive Field Network for Real-Time Driving Scene Semantic Segmentation [J].
Li, Dehui ;
Cao, Zhiguo ;
Xian, Ke ;
Yang, Jiaqi ;
Qi, Xinyuan ;
Li, Wei .
PRICAI 2019: TRENDS IN ARTIFICIAL INTELLIGENCE, PT III, 2019, 11672 :350-362
[47]   BMSeNet: Multiscale Context Pyramid Pooling and Spatial Detail Enhancement Network for Real-Time Semantic Segmentation [J].
Zhao, Shan ;
Zhao, Xin ;
Huo, Zhanqiang ;
Zhang, Fukai .
SENSORS, 2024, 24 (16)
[48]   Real-time semantic segmentation based on improved BiSeNet [J].
Ren F. ;
Yang L. ;
Zhou H. ;
Zhang S. ;
He X. ;
Xu W. .
Guangxue Jingmi Gongcheng/Optics and Precision Engineering, 2023, 31 (08) :1217-1227
[49]   A Lightweight Road Scene Semantic Segmentation Algorithm [J].
Peng, Jiansheng ;
Yang, Qing ;
Hou, Yaru .
CMC-COMPUTERS MATERIALS & CONTINUA, 2023, 77 (02) :1929-1948
[50]   Lookahead adversarial learning for near real-time semantic segmentation [J].
Jamali-Rad, Hadi ;
Szabo, Attila .
COMPUTER VISION AND IMAGE UNDERSTANDING, 2021, 212