IDD-YOLOv7：一种用于输电线路绝缘子多缺陷的轻量化检测方法

doi:10.11996/JG.j.2095-302X.2024010090

图学学报 ›› 2024, Vol. 45 ›› Issue (1): 90-101.DOI: 10.11996/JG.j.2095-302X.2024010090

• 图像处理与计算机视觉 • 上一篇下一篇

IDD-YOLOv7：一种用于输电线路绝缘子多缺陷的轻量化检测方法

翟永杰(), 赵晓瑜, 王璐瑶, 王亚茹(), 宋晓轲, 朱浩硕

华北电力大学自动化系，河北省保定市 071003

收稿日期:2023-08-11 接受日期:2023-10-31 出版日期:2024-02-29 发布日期:2024-02-29
通讯作者:王亚茹(1990-)，女，讲师，博士。主要研究方向为模式识别、数据挖掘、输电线路元件检测等。E-mail：wangyaru@ncepu.edu.cn
第一作者:翟永杰(1972-)，男，教授，博士。主要研究方向为电力视觉。E-mail：zhaiyongjie@ncepu.edu.cn
基金资助:
国家自然科学基金联合基金重点支持项目(U21A20486);中央高校基本科研业务费专项资金项目(2023JC006);国家自然科学基金青年基金项目(62303184);河北省自然科学基金青年科学基金项目(F2021502008)

IDD-YOLOv7: a lightweight method for multiple defect detection of insulators in transmission lines

ZHAI Yongjie(), ZHAO Xiaoyu, WANG Luyao, WANG Yaru(), SONG Xiaoke, ZHU Haoshuo

Department of Automation, North China Electric Power University, Baoding Hebei 071003, China

Received:2023-08-11 Accepted:2023-10-31 Published:2024-02-29 Online:2024-02-29
First author：ZHAI Yongjie (1972-), professor, Ph.D. His main research interest covers power vision. E-mail：zhaiyongjie@ncepu.edu.cn
Supported by:
National Natural Science Foundation of China Joint Fund Project Key Support Project(U21A20486);Fundamental Research Funds for the Central Universities(2023JC006);National Natural Science Foundation Youth Fund project(62303184);Natural Science Foundation Youth Science Fund Project of Hebei Province(F2021502008)

摘要/Abstract

摘要：

YOLO目标检测算法是当前基于图像的输电线路绝缘子缺陷检测的主流方法，然而现有模型复杂度较大，亟需合理有效的参数压缩方法作为前提条件，来为解决无人机边缘设备部署的困境问题奠定基础；同时，无人机航拍的绝缘子缺陷图像背景复杂、缺陷尺寸较小，容易出现误检、漏检等问题。为此，提出了一种用于输电线路绝缘子多缺陷检测的Insulator Defect Detection-YOLOv7(IDD-YOLOv7)模型，以降低模型复杂度，提高模型鲁棒性。首先，在多尺度特征融合的过程中加入坐标注意力(Coordinate Attention)机制，抑制复杂背景的干扰，提升模型对小目标的全局感知能力；之后，设计C3GhostNetV2模块，用于捕获不同空间像素之间的远程依赖性，在增强模型表达能力的同时降低模型的参数量和浮点运算量；最后，提出Focal-CIoU损失函数，提高模型高质量anchor的贡献，加快模型的收敛速度。实验结果表明，本文方法与基线模型相比mAP⁵⁰提升了3.8%，查准率和召回率分别提升了1.7%和7.6%，参数量和浮点运算量分别下降了18.3%和14.0%，绝缘子自爆、破损、闪络缺陷的AP⁵⁰分别提升了0.8%、4.5%、6.3%。

华北电力大学翟永杰教授及其学生赵晓瑜等针对输电线路绝缘子缺陷图像背景复杂、缺陷尺寸较小且模型复杂度较大等问题，引入坐标注意力机制，提升模型对小目标的全局感知能力；设计C3GhostNetV2模块，增强模型表达能力的同时降低模型的参数量和浮点运算量；提出Focal-CIoU损失函数，提高模型高质量anchor的贡献，加快模型的收敛速度。实验结果表明，本文方法提高了模型检测精度，降低了模型复杂度，为解决无人机边缘设备部署问题奠定基础。

关键词: YOLOv7, 绝缘子缺陷检测, 注意力机制, 模型复杂度, 轻量化, 损失函数

Abstract:

The YOLO objective detection algorithm is currently the mainstream method for detecting insulator defects in image-based power transmission lines. However, due to the high complexity of existing models, a reasonable and effective parameter compression method is urgently needed as a prerequisite to establish the foundation for solving the dilemma of UAV edge device deployment. Additionally, the complex background of the insulator defect images captured by drones and small size of defects can lead to problems such as false detections and omissions. To address these issues, the Insulator Defect Detection-YOLOv7 (IDD-YOLOv7) model was proposed for multi-defect detection in power transmission line insulators, aiming to reduce model complexity and enhance robustness. Firstly, a coordinate attention mechanism was incorporated during the multi-scale feature fusion process to suppress interference from complex backgrounds and enhance the model’s global perception of small objects. Secondly, a C3GhostNetV2 module was designed to capture long-range dependencies between different spatial pixels, thus enhancing the model’s expressive power while reducing the parameter quantity and floating-point operation complexity. Lastly, the Focal-CIoU loss function was proposed to improve the contribution of high-quality anchors to the model and accelerate model convergence. Experimental results demonstrated that compared with the baseline model, the mAP⁵⁰ of this method has increased by 3.8%, with precision and recall rates increasing by 1.7% and 7.6%, respectively, and the parameter quantity and floating-point operations have decreased by 18.3% and 14.0%, respectively. The AP⁵⁰ of insulator self-explosion, damage, and flashover defects have increased by 0.8%, 4.5%, and 6.3%, respectively.

Key words: YOLOv7, insulator defect detection, attention mechanism, model complexity, lightweight, loss function

中图分类号:

TP391

翟永杰, 赵晓瑜, 王璐瑶, 王亚茹, 宋晓轲, 朱浩硕. IDD-YOLOv7：一种用于输电线路绝缘子多缺陷的轻量化检测方法[J]. 图学学报, 2024, 45(1): 90-101.

ZHAI Yongjie, ZHAO Xiaoyu, WANG Luyao, WANG Yaru, SONG Xiaoke, ZHU Haoshuo. IDD-YOLOv7: a lightweight method for multiple defect detection of insulators in transmission lines[J]. Journal of Graphics, 2024, 45(1): 90-101.

图/表 20

图1 IDD-YOLOv7网络结构

Fig. 1 IDD-YOLOv7 network structure

图2 CA模块结构

Fig. 2 CA module structure

图3 C3GhostNetV2结构

Fig. 3 C3GhostNetV2 structure

图4 GhostNetV2 bottleneck结构

Fig. 4 GhostNetV2 bottleneck structure

表1 实验硬件环境

Table 1 Experimental hardware environment

硬件名称	型号	数量
CPU	Intel(R) Core(TM) i9-10850K	1
内存	金士顿 16 G DDR4	2
显卡	GeForce RTX 2080ti	1
硬盘	西数10 TB	1

图5 各类缺陷样本图((a)自爆；(b)破损；(c)闪络)

Fig. 5 Samples of various defects ((a) Drope; (b) Damage; (c) Flashover)

表2 不同注意力性能比较

Table 2 Comparison of different attention performances

模型	AP⁵⁰ 自爆	AP⁵⁰ 破损	AP⁵⁰ 闪络	mAP⁵⁰/ %	Param/ M	GFLOPs/ G
SE	95.3	78.5	77.9	83.9	37.9	105.7
CBAM	97.0	76.2	77.4	83.5	37.3	105.2
CA	97.2	80.2	76.8	84.7	37.2	105.2

表3 不同超参数性能比较

Table 3 Performance comparison of different hyperparameters

超参数γ	AP⁵⁰ 自爆	AP⁵⁰ 破损	AP⁵⁰ 闪络	mAP⁵⁰/ %
0	96.6	78.0	71.9	82.2
0.25	96.5	76.8	75.3	82.7
0.5	97.4	78.2	76.4	84.0
0.75	97.6	76.9	75.8	83.4
1	96.8	77.1	74.6	82.8

表4 YOLOv7改进模型的消融实验结果比较

Table 4 Comparison of ablation experiment results of YOLOv7 improved model

实验	C3GhostNetV2	CA	Focal-CIoU	AP⁵⁰ 自爆	AP⁵⁰ 破损	AP⁵⁰ 闪络	mAP⁵⁰/ %	查准率/ %	召回率/ %	参数量/ M	GFLOPs/ G
1				96.6	78.0	71.9	82.2	86.7	76.6	37.2	105.1
2	√			96.8	77.4	74.6	82.9	87.3	83.3	30.3	90.3
3		√		97.2	80.2	76.8	84.7	87.2	83.1	37.2	105.2
4			√	97.4	78.2	76.4	84.0	88.7	83.2	37.2	105.1
5	√	√		96.9	79.9	77.5	84.8	88.1	82.8	30.4	90.4
6	√	√	√	97.4	82.5	78.2	86.0	88.4	84.2	30.4	90.4

图6 热力图可视化对比结果((a)自爆场景1；(b)自爆场景2；(c)破损场景1；(d)破损场景2；(e)闪络场景1；(f)闪络场景2)

Fig. 6 Heat map visual comparison results ((a) Drope scenario 1; (b) Drope scenario 2; (c) Damage scenario 1; (d) Damage scenario 2; (e) Flashover scenario 1; (f) Fashover scenario 2)

图7 不同损失函数下的mAP50数据曲线

Fig. 7 mAP50 data curves under different loss functions

图8 不同损失函数下的mAP50:95数据曲线

Fig. 8 mAP50:95 data curves under different loss functions

表5 不同损失性能比较

Table 5 Comparison of different loss performance

损失函数	AP⁵⁰ 自爆	AP⁵⁰ 破损	AP⁵⁰ 闪络	mAP⁵⁰/ %	mAP^50:95/ %
Focal-DIoU	96.5	79.0	73.2	82.9	50.7
Focal-GIoU	96.6	79.8	72.7	83.0	49.6
Focal-EIoU	97.2	77.1	76.1	83.5	51.2
Focal-CIoU	97.4	78.2	76.4	84.0	51.5

表6 不同模型性能比较

Table 6 Performance comparison of different models

检测算法	AP⁵⁰ 自爆	AP⁵⁰ 破损	AP⁵⁰ 闪络	mAP⁵⁰/ %	mAP^50:95/ %
Faster R-CNN	91.2	64.5	69.7	75.2	49.3
SSD	84.6	54.3	35.4	58.1	44.5
YOLOv5	96.8	74.6	77.2	82.9	51.8
GBH-YOLOv5	96.3	79.9	73.0	83.1	50.5
TPH-YOLOv5	97.1	79.5	76.0	84.2	51.1
YOLOv7	96.6	78.0	71.9	82.2	51.1
YOLOv8	96.5	78.5	72.7	82.6	51.7
本文算法	97.4	82.5	78.2	86.0	52.8

图9 不同模型可视化对比结果((a)自爆场景1；(b)自爆场景2；(c)破损场景1；(d)破损场景2；(e)闪络场景1；(f)闪络场景2)

Fig. 9 Visual comparison results of different models ((a) Drope scenario 1; (b) Drope scenario 2; (c) Damage scenario 1; (d) Damage scenario 2; (e) Flashover scenario 1; (f) Flashover scenario 2)

表7 不同模型在不同环境的性能比较/%

Table 7 Performance comparison of different models in different environments/%

检测算法	正常图像mAP⁵⁰	暗化图像mAP⁵⁰	增亮图像mAP⁵⁰	有雾图像mAP⁵⁰	有雨图像mAP⁵⁰
Faster R-CNN	75.2	66.1	68.3	69.8	67.0
SSD	58.1	54.5	53.0	50.9	51.9
YOLOv5	82.9	81.8	78.2	81.5	81.4
GBH-YOLOv5	83.1	81.1	78.9	80.2	81.2
TPH-YOLOv5	84.2	82.9	79.4	82.2	82.4
YOLOv7	82.2	80.7	78.3	80.9	80.8
YOLOv8	82.6	81.3	79.2	81.2	81.7
本文算法	86.0	84.9	83.3	84.9	85.3

表8 不同模型在不同环境mAP50下降百分比/%

Table 8 Different models in different environments mAP50 drop percentage/%

检测算法	暗化图像mAP⁵⁰	增亮图像mAP⁵⁰	有雾图像mAP⁵⁰	有雨图像mAP⁵⁰
Faster R-CNN	12.1	9.2	7.2	10.9
SSD	6.2	8.8	12.4	10.7
YOLOv5	1.3	5.7	1.7	1.8
GBH-YOLOv5	2.4	5.1	3.5	2.3
TPH-YOLOv5	1.5	5.7	2.4	2.1
YOLOv7	1.8	4.7	1.6	1.7
YOLOv8	1.6	4.1	1.7	1.1
本文算法	1.3	3.1	1.3	0.8

图10 模拟不同自然环境的可视化结果((a)自爆场景1；(b)自爆场景2；(c)破损场景1；(d)破损场景2；(e)闪络场景1；(f)闪络场景2)

Fig. 10 Visualization results of simulated different natural environments ((a) Drope scenario 1; (b) Drope scenario 2; (c) Damage scenario 1; (d) Damage scenario 2; (e) Flashover scenario 1; (f) Flashover scenario 2)

表9 不同模型在RSOD数据集性能比较

Table 9 Comparison of performance of different models on RSOD dataset

检测算法	AP⁵⁰ 飞机	AP⁵⁰ 油桶	AP⁵⁰ 立交桥	AP⁵⁰ 操场	mAP⁵⁰/ %
Faster R-CNN	92.1	97.4	74.2	99.3	90.8
SSD	75.1	95.1	80.8	99.6	87.7
YOLOv5	94.5	97.4	91.0	99.2	95.5
GBH-YOLOv5	94.6	96.9	89.3	99.3	95.0
TPH-YOLOv5	94.4	97.5	93.1	99.2	96.1
YOLOv7	94.0	96.8	93.8	98.9	95.9
YOLOv8	93.4	97.0	93.7	98.7	95.7
本文算法	95.5	98.5	95.3	99.8	97.3

图11 RSOD数据集可视化结果 ((a)飞机；(b)油桶；(c)立交桥；(d)操场)

Fig. 11 RSOD dataset visualization results ((a) Aircraft; (b) Oiltank; (c) Overpass; (d) Playground)

参考文献 30

[1]	OGBONNA V E, POPOOLA P I, POPOOLA O M, et al. A comparative study on the failure analysis of field failed high voltage composite insulator core rods and recommendation of composite insulators: a review[J]. Engineering Failure Analysis, 2022, 138: 106369. DOI URL
[2]	张新春, 曹应平, 韩春雨, 等. 基于图像处理的输电线路导线表面损伤特征研究[J]. 图学学报, 2018, 39(3): 440-447. DOI
	ZHANG X C, CAO Y P, HAN C Y, et al. On the surface damage features of transmission line conductors based on image processing[J]. Journal of Graphics, 2018, 39(3): 440-447 (in Chinese). DOI
[3]	赵振兵, 蒋志钢, 李延旭, 等. 输电线路部件视觉缺陷检测综述[J]. 中国图象图形学报, 2021, 26(11): 2545-2560.
	ZHAO Z B, JIANG Z G, LI Y X, et al. Overview of visual defect detection of transmission line components[J]. Journal of Image and Graphics, 2021, 26(11): 2545-2560 (in Chinese).
[4]	PARK K C, MOTAI Y, YOON J R. Acoustic fault detection technique for high-power insulators[J]. IEEE Transactions on Industrial Electronics, 2017, 64(12): 9699-9708. DOI URL
[5]	高文婷, 刘越. 面向移动增强现实的实时深度学习目标检测方法综述[J]. 图学学报, 2021, 42(4): 525-534.
	GAO W T, LIU Y. Review of real-time deep learning-based object detection for mobile augmented reality[J]. Journal of Graphics, 2021, 42(4): 525-534 (in Chinese).
[6]	冯晗, 姜勇. 使用改进Yolov5的变电站绝缘子串检测方法[J]. 智能系统学报, 2023, 18(2): 325-332.
	FENG H, JIANG Y. A substation insulator string detection method based on an improved Yolov5[J]. CAAI Transactions on Intelligent Systems, 2023, 18(2): 325-332 (in Chinese).
[7]	贾晓芬, 于业齐, 郭永存, 等. 航拍绝缘子自爆缺陷的轻量化检测方法[J]. 高电压技术, 2023, 49(1): 294-300.
	JIA X F, YU Y Q, GUO Y C, et al. Lightweight detection method of self-explosion defect of aerial photo insulator[J]. High Voltage Engineering, 2023, 49(1): 294-300 (in Chinese).
[8]	LI X F, SU H S, LIU G H. Insulator defect recognition based on global detection and local segmentation[J]. IEEE Access, 2020, 8: 59934-59946. DOI URL
[9]	REN S Q, HE K M, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149. DOI PMID
[10]	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. Cham: Springer, 2015: 234-241.
[11]	TAO X, ZHANG D P, WANG Z H, et al. Detection of power line insulator defects using aerial images analyzed with convolutional neural networks[J]. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2020, 50(4): 1486-1498. DOI URL
[12]	LIU J J, LIU C Y, WU Y Q, et al. Insulators' identification and missing defect detection in aerial images based on cascaded YOLO models[J]. Computational Intelligence and Neuroscience, 2022, 2022: 7113765.
[13]	ZHAO Z B, ZHEN Z, ZHANG L, et al. Insulator detection method in inspection image based on improved faster R-CNN[J]. Energies, 2019, 12(7): 1204. DOI URL
[14]	宋智伟, 黄新波, 纪超, 等. 基于Flexible YOLOv7的输电线路绝缘子缺陷检测和故障预警方法[J]. 高电压技术, 2023, 49(12): 5084-5094.
	SONG Z W, HUANG X B, JI C, et al. Insulator defect detection and fault warning method for transmisson line based on Flexible YOLOv7[J]. High Voltage Engineering, 2023, 49(12): 5084-5094 (in Chinese).
[15]	翟永杰, 杨珂, 王乾铭, 等. 基于混合样本迁移学习的盘型绝缘子缺陷检测[J]. 中国电机工程学报, 2023, 43(7): 2867-2877.
	ZHAI Y J, YANG K, WANG Q M, et al. Disc insulator defect detection based on mixed sample transfer learning[J]. Proceedings of the CSEE, 2023, 43(7): 2867-2877 (in Chinese).
[16]	WANG C Y, BOCHKOVSKIY A, LIAO H Y M. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2023: 7464-7475.
[17]	HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2018: 7132-7141.
[18]	WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]// European Conference on Computer Vision. Cham: Springer, 2018: 3-19.
[19]	HOU Q B, ZHOU D Q, FENG J S. Coordinate attention for efficient mobile network design[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2021: 13708-13717.
[20]	CHEN B Y, DANG Z C. Fast PCB defect detection method based on FasterNet backbone network and CBAM attention mechanism integrated with feature fusion module in improved YOLOv7[J]. IEEE Access, 2023, 11: 95092-95103. DOI URL
[21]	TANG Y H, HAN K, GUO J Y, et al. GhostNetV2: enhance cheap operation with long-range attention[EB/OL]. [2023-06-19]. https://arxiv.org/abs/2211.12905.pdf.
[22]	ZHENG Z H, WANG P, LIU W, et al. Distance-IoU loss: faster and better learning for bounding box regression[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(7): 12993-13000. DOI URL
[23]	LIU S J, ZHOU H B, LI C M, et al. Analysis of anchor-based and anchor-free object detection methods based on deep learning[C]// 2020 IEEE International Conference on Mechatronics and Automation. New York: IEEE Press, 2020: 1058-1065.
[24]	LIU Y, LIU D, HUANG X, et al. Insulator defect detection with deep learning: A survey[J]. IET Generation, Transmission & Distribution, 2023, 17(16): 3541-3558.. DOI URL
[25]	SELVARAJU R R, COGSWELL M, DAS A, et al. Grad-CAM: visual explanations from deep networks via gradient-based localization[C]// 2017 IEEE International Conference on Computer Vision. New York: IEEE Press, 2017: 618-626.
[26]	REZATOFIGHI H, TSOI N, GWAK J, et al. Generalized intersection over union: a metric and a loss for bounding box regression[C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2020: 658-666.
[27]	ZHANG Y F, REN W Q, ZHANG Z, et al. Focal and efficient IOU loss for accurate bounding box regression[J]. Neurocomputing, 2022, 506: 146-157. DOI URL
[28]	LI L L, WANG Z F, ZHANG T T. GBH-YOLOv5: ghost convolution with BottleneckCSP and tiny target prediction head incorporating YOLOv5 for PV panel defect detection[J]. Electronics, 2023, 12(3): 561. DOI URL
[29]	ZHU X K, LYU S C, WANG X, et al. TPH-YOLOv5: improved YOLOv5 based on transformer prediction head for object detection on drone-captured scenarios[C]// 2021 IEEE/CVF International Conference on Computer Vision Workshops. New York: IEEE Press, 2021: 2778-2788.
[30]	LI K, WAN G, CHENG G, et al. Object detection in optical remote sensing images: a survey and a new benchmark[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2020, 159: 296-307. DOI URL

IDD-YOLOv7：一种用于输电线路绝缘子多缺陷的轻量化检测方法

IDD-YOLOv7: a lightweight method for multiple defect detection of insulators in transmission lines

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 20

参考文献 30

相关文章 15

编辑推荐

Metrics

本文评价

[1]	李大湘, 吉展, 刘颖, 唐垚. 改进YOLOv7遥感图像目标检测算法[J]. 图学学报, 2024, 45(4): 650-658.
[2]	魏敏, 姚鑫. 基于多尺度与注意力机制的两阶段风暴单体外推研究[J]. 图学学报, 2024, 45(4): 696-704.
[3]	曾志超, 徐玥, 王景玉, 叶元龙, 黄志开, 王欢. 基于SOE-YOLO轻量化的水面目标检测算法[J]. 图学学报, 2024, 45(4): 736-744.
[4]	武兵, 田莹. 基于注意力机制的多尺度道路损伤检测算法研究[J]. 图学学报, 2024, 45(4): 770-778.
[5]	赵磊, 李栋, 房建东, 曹琪. 面向交通标志的改进YOLO目标检测算法[J]. 图学学报, 2024, 45(4): 779-790.
[6]	朱强军, 胡斌, 汪慧兰, 王杨. 基于轻量化YOLOv8s交通标志的检测[J]. 图学学报, 2024, 45(3): 422-432.
[7]	李跃华, 仲新, 姚章燕, 胡彬. 基于改进YOLOv5s的着装不规范检测算法研究[J]. 图学学报, 2024, 45(3): 433-445.
[8]	张相胜, 杨骁. 基于改进YOLOv7-tiny的橡胶密封圈缺陷检测方法[J]. 图学学报, 2024, 45(3): 446-453.
[9]	李滔, 胡婷, 武丹丹. 结合金字塔结构和注意力机制的单目深度估计[J]. 图学学报, 2024, 45(3): 454-463.
[10]	黄友文, 林志钦, 章劲, 陈俊宽. 结合坐标Transformer的轻量级人体姿态估计算法[J]. 图学学报, 2024, 45(3): 516-527.
[11]	路龙飞, 王峻峰, 赵世闻, 李广, 丁鑫涛. 基于力位感知技能学习的轴孔柔顺装配方法[J]. 图学学报, 2024, 45(2): 250-258.
[12]	郭宗洋, 刘立东, 蒋东华, 刘子翔, 朱熟康, 陈京华. 基于语义引导神经网络的人体动作识别算法[J]. 图学学报, 2024, 45(1): 26-34.
[13]	苑朝, 赵亚冬, 张耀, 王嘉璇, 徐大伟, 翟永杰, 朱松松. 基于YOLO轻量化的多模态行人检测算法[J]. 图学学报, 2024, 45(1): 35-46.
[14]	古天骏, 熊苏雅, 林晓. 基于SASGAN的戏剧脸谱多样化生成[J]. 图学学报, 2024, 45(1): 102-111.
[15]	崔克彬, 焦静颐. 基于MCB-FAH-YOLOv8的钢材表面缺陷检测算法[J]. 图学学报, 2024, 45(1): 112-125.