Regarding the issue that existing semantic segmentation algorithms fail to fully exploit target edge features in autonomous driving scenarios, an edge feature-guided semantic segmentation method for intelligent vehicles was proposed. First, an edge feature enhancement module was introduced to extract edge information from the original image as a parallel input to the encoder, strengthening the neural network's edge feature representation. Second, an edge attention module was proposed to fuse original image features and edge features, adaptively balancing their contributions in different regions for semantic segmentation. Finally, an efficient channel pyramid pooling module was designed in the decoder to improve the utilization efficiency of contextual semantic features after pooling at different scales. Experimental results on the Cityscapes dataset demonstrate that the proposed method achieves 76.1% MIoU with a processing speed of 132.6 f/s, outperforming the baseline PP-LiteSeg by 2.2% MIoU while maintaining real-time performance, balancing segmentation accuracy and computational efficiency.
注意力机制通过对特征的不同位置给予不同的权重来关注重点区域特征[19]。传统自注意力模块的Key张量、Query张量和Value张量( Q 、 K 、 V )均来源于相同特征图。与传统自注意力机制不同,本文边缘注意力机制的Key张量来源于边缘特征,Query张量和Value张量来源于原始图像特征。边缘注意力模块通过借助边缘特征,强调原始图像特征中的特定区域,其利用点积计算边缘特征张量Key与原始图像特征张量Query之间的相似度,并利用Softmax函数获取归一化权值,其计算公式如下所示。
为了评估本文算法的性能,参考当前主流的语义分割评估方法[22,23],本文采用平均交并比(Mean intersection over union,MIoU)评估语义分割算法精度,采用算法推理速度(Frames per second,FPS)评估语义分割算法实时性。其中,FPS为单位时间(1 s)内处理的图像帧数;MIoU为像素级别下的预测与真实标注的平均重叠程度,MIoU的计算公式可表示为:
MinLi, DongBing-jie, AnDong. Semantic segmentation algorithm based on multi-attention mechanism and cross-feature fusion[J]. Computer Engineering, 2024, 50(8): 282-289.
[3]
LiangY, XuZ, RastiS, et al. On the use of a semantic segmentation micro-service in AR devices for UI placement[C]∥2022 IEEE Games, Entertainment, Media Conference, Bridgetown, Barbados, 2022: 1-6.
[4]
YangL, BaiY, RenF, et al. Lcfnets: compensation strategy for real-time semantic segmentation of autonomous driving[J]. IEEE Transactions on Intelligent Vehicles, 2024, 9(4): 4715-4729.
[5]
ZhouZ, Rahman SiddiqueeM M, TajbakhshN, et al. Unet++: a nested U-net architecture for medical image segmentation[C]∥Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support, Granada, Spain, 2018: 3-11.
[6]
MoY, WuY, YangX, et al. Review the state-of-the-art technologies of semantic segmentation based on deep learning[J]. Neurocomputing, 2022, 493: 626-646.
[7]
MuhammadK, HussainT, UllahH, et al. Vision-based semantic segmentation in scene understanding for autonomous driving: recent achievements, challenges, and outlooks[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(12): 22694-22715.
[8]
ShelhamerE, LongJ, DarrellT. Fully convolutional networks for semantic segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 39(4): 640-651.
[9]
RonnebergerO, FischerP, BroxT. U-net: convolutional networks for biomedical image segmentation[C]∥Medical Image Computing and Computer-Assisted Intervention, Munich, Germany, 2015: 234-241.
[10]
ChenL C, PapandreouG, KokkinosI, et al. Deeplab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 40(4): 834-848.
[11]
ZhaoH S, ShiJ P, QiX J, et al. Pyramid scene parsing network[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,Honolulu, USA, 2017: 6230-6239,.
[12]
PaszkeA, ChaurasiaA, KimS, et al. Enet: a deep neural network architecture for real-time semantic segmentation[J/OL].[2025-05-10].
[13]
YuC, WangJ, PengC, et al. BiSeNet: bilateral segmentation network for real-time semantic segmentation[C]∥Proceedings of the European Conference on Computer Vision (ECCV),Munich, Germany, 2018: 334-349.
[14]
PengJ, LiuY, TangS, et al. Pp-liteseg: a superior real-time semantic segmentation model[J/OL]. [2025-05-10].
[15]
XuJ, XiongZ, BhattacharyyaS P. PIDNet: a real-time semantic segmentation network inspired by PID controllers[C]∥Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, Canada,2023: 19529-19539.
[16]
TakikawaT, AcunaD, JampaniV, et al. Gated-SCNN: Gated shape cnns for semantic segmentation[C]∥Proceedings of the IEEE/CVF International Conference on Computer Vision, Seoul, Korea, 2019: 5228-5237.
[17]
YuL, YaoA, DuanJ. Improving semantic segmentation via decoupled body and edge information[J]. Entropy, 2023, 25(6): No. 891.
[18]
DingL, GoshtasbyA. On the Canny edge detector[J]. Pattern Recognition, 2001, 34(3): 721-725.
[19]
FanM, LaiS, HuangJ, et al. Rethinking bisenet for real-time semantic segmentation[C]∥Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,Nashville, USA, 2021: 9711-9720.
[20]
NiuZ, ZhongG, YuH. A review on the attention mechanism of deep learning[J]. Neurocomputing, 2021, 452: 48-62.
[21]
WangQ, WuB, ZhuP, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]∥Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, USA, 2020: 11531-11539.
[22]
CordtsM, OmranM, RamosS, et al. The cityscapes dataset for semantic urban scene understanding[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,Las Vegas, USA, 2016: 3213-3223.
[23]
ZhouQ, WangL, GaoG, et al. Boundary-guided lightweight semantic segmentation with multi-scale semantic context[J]. IEEE Transactions on Multimedia, 2024, 26: 7887-7900.
PerazziF, Pont-TusetJ, McWilliamsB, et al. A benchmark dataset and evaluation methodology for video object segmentation[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA, 2016: 724-732.
[26]
YuC, GaoC, WangJ, et al. Bisenet v2: bilateral network with guided aggregation for real-time semantic segmentation[J]. International Journal of Computer Vision, 2021, 129: 3051-3068.
[27]
HanH Y, ChenY C, HsiaoP Y, et al. Using channel-wise attention for deep CNN based real-time semantic segmentation with class-aware edge information[J]. IEEE Transactions on Intelligent Transportation Systems, 2020, 22(2): 1041-1051.
[28]
TsaiT H, TsengY W. BiSeNet V3: bilateral segmentation network with coordinate attention for real-time semantic segmentation[J]. Neurocomputing, 2023, 532: 33-42.
[29]
LiuQ, WangC, LiZ, et al. Attention based lightweight asymmetric network for real-time semantic segmentation[J]. Engineering Applications of Artificial Intelligence, 2024, 130: No.107736.