A Multi-scale Edge Detection Method Based on Encoder-Decoder

doi:10.12068/j.issn.1005-3026.2024.07.004

Abstract

Abstract:

Traditional edge detection algorithms often suffer from fractured and discontinuous edge lines， while deep learning?based approaches fail to ensure edge clarity and accuracy， often leading to the loss of edge details. To address these issues， a lightweight edge detection method based on encoder?decoder structure is proposed. The encoder extracts the edge features of the image， recovers the edge information lost when sampling under the encoder at the decoding end， uses the jump connection method between the codecs to achieve the fusion between low?level features and high?level features， and uses the deep supervision module with attention mechanism to further learn multi?scale and multi?level edge features to generate fine image edges. The network model is trained on the BSDS500S dataset. Experimental results show that the ODS and OIS of the proposed method reach 0.808 and 0.830 respectively， and the frame rate on the GTX 1060 machine reaches 60 frames/second， which exceeds the mainstream edge detection methods based on convolutional neural networks， thus showing its effectiveness.

Key words: edge detection, encoder?decoder, skip connection, attention mechanism, deep supervision

CLC Number:

TP 391

An-lin TIAN, Wei-min LEI, Peng ZHANG, Wei ZHANG. A Multi-scale Edge Detection Method Based on Encoder-Decoder[J]. Journal of Northeastern University(Natural Science), 2024, 45(7): 936-943.

Figures/Tables 13

References 23

1	Zhang M H， Luo H L， Song W，et al.Spectral‑spatial offset graph convolutional networks for hyperspectral image classification［J］.Remote Sensing，2021，13（21）：4342.
2	王玲敏，段军，辛立伟.引入注意力机制的YOLOv5安全帽佩戴检测方法［J］.计算机工程与应用，2022，58（9）：303-312.
	Wang Ling‑min， Duan Jun， Xin Li‑wei.Detection method of YOLOv5 helmet wearing with attention mechanism［J］.Computer Engineering and Application，2022，58（9）：303-312.
3	Wang R， Qiu K.Fine‑grained remote sensing semantic segmentation method under multi‑stage supervision learning［C］//14th International Conference on Measuring Technology and Mechatronics Automation（ICMTMA）.Changsha，2022：361-365.
4	Sobel I， Feldman G.A 3x3 isotropic gradient operator for image processing［EB/OL］.（2015-06-14）［2022-11-10］..
5	Prewitt J M S.Object enhancement and extraction［J］.Picture Processing and Psychopictorics，1970，10（1）：15-19.
6	Canny J.A computational approach to edge detection［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，1986，8（6）：679-698.
7	Arbelaez P， Maire M， Fowlkes C，et al.Contour detection and hierarchical image segmentation［J］.IEEE transactions on Pattern Analysis and Machine Intelligence，2010，33（5）：898-916.
8	Dollár P， Zitnick C L.Fast edge detection using structured forests［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，2014，37（8）：1558-1570.
9	刘胜男，宁纪锋.基于超像素的点互信息边界检测算法［J］.计算机应用，2016，36（8）：2296-2300.
	Liu Sheng‑nan， Ning Ji‑feng.Point mutual information boundary detection algorithm based on super pixel［J］.Computer Applications，2016，36（8）：2296-2300.
10	Shen W， Wang X G， Wang Y，et al.Deepcontour：a deep convolutional feature learned by positive‑sharing loss for contour detection［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Boston，2015：3982-3991.
11	Bertasius G， Shi J， Torresani L.Deepedge：a multi‑scale bifurcated deep network for top‑down contour detection［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Boston，2015：4380-4389.
12	Ganin Y， Lempitsky V.N4-Fields：neural network nearest neighbor fields for image transforms［C］//Asian Conference on Computer Vision.Cham：Springer，2014：536-551.
13	Xie S N， Tu Z W.Holistically‑nested edge detection［C］//Proceedings of the IEEE International Conference on Computer Vision.Santiago，2015：1395-1403.
14	Simonyan K， Zisserman A.Very deep convolutional networks for large‑scale image recognition［EB/OL］.（2015-04-10）［2022-11-10］..
15	Liu Y， Cheng M M， Hu X，et al.Richer convolutional features for edge detection［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Honolalu，2017：3000-3009.
16	Wang Y P， Zhao X， Huang K Q.Deep crisp boundaries［C］//Proceedings of the IEEE Conference On Computer Vision and Pattern Recognition.Honolulu，2017：3892-3900.
17	He J Z， Zhang S L， Yang M，et al.Bi‑directional cascade network for perceptual edge detection［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Long Beach，2019：3828-3837.
18	Deng R X， Shen C H， Liu S J，et al.Learning to predict crisp boundaries［C］//Proceedings of the European Conference on Computer Vision（ECCV）.Munich：Springer，2018：562-578.
19	Su Z， Liu W Z， Yu Z T，et al.Pixel difference networks for efficient edge detection［C］//Proceedings of the IEEE/CVF International Conference on Computer Vision.Montreal，2021：5117-5127.
20	Badrinarayanan V， Kendall A， Cipolla R.SegNet：a deep convolutional encoder‑decoder architecture for image segmentation［J］.IEEE Transactions on Pattern Analysis & Machine Intelligence，2015，39：2481-2495.
21	Ronneberger O， Fischer P， Brox T.U‑net：convolutional networks for biomedical image segmentation［C］//International Conference on Medical Image Computing and Computer‑Assisted Intervention.Munich：Springer，2015：234-241.
22	Howard A， Sandler M， Chu G，et al.Searching for MobileNetV3［C］//Proceedings of the IEEE/CVF International Conference on Computer Vision.Seoul，2019：1314-1324.
23	Hallman S， Fowlkes C C.Oriented edge forests for boundary detection［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Boston，2015：1732-1740.

方法	ODS	OIS	AP
Human	0.803	0.803	—
Canny^[6]	0.611	0.676	0.520
Sobel^[4]	0.539	0.575	0.498
gPb-UCM^[7]	0.729	0.755	0.745
SE^[8]	0.743	0.764	0.800
OEF（oriented edge forests）^[23]	0.746	0.770	0.815
DeepContour^[10]	0.757	0.776	0.790
DeepEdge^[11]	0.753	0.772	0.807
HED^[13]	0.788	0.808	0.815
RCF^[15]	0.806	0.823	0.834
本文	0.808	0.830	0.843

方法	ODS	OIS	AP
Human	0.803	0.803	—
Canny^[6]	0.611	0.676	0.520
Sobel^[4]	0.539	0.575	0.498
gPb-UCM^[7]	0.729	0.755	0.745
SE^[8]	0.743	0.764	0.800
OEF（oriented edge forests）^[23]	0.746	0.770	0.815
DeepContour^[10]	0.757	0.776	0.790
DeepEdge^[11]	0.753	0.772	0.807
HED^[13]	0.788	0.808	0.815
RCF^[15]	0.806	0.823	0.834
本文	0.808	0.830	0.843

方法	ODS	OIS	AP
gPb-UCM^[7]	0.631	0.661	0.562
OEF^[23]	0.651	0.667	0.653
SE^[8]	0.695	0.708	0.719
HED^[13]	0.741	0.757	0.749
本文	0.753	0.769	0.763

方法	ODS	OIS	AP
gPb-UCM^[7]	0.631	0.661	0.562
OEF^[23]	0.651	0.667	0.653
SE^[8]	0.695	0.708	0.719
HED^[13]	0.741	0.757	0.749
本文	0.753	0.769	0.763

配置	ODS	OIS
无CAM	0.792	0.809
Encoder-CAM	0.800	0.815
Decoder-CAM	0.808	0.830