基于关键点检测的服装尺寸测量方法

Clothing size measurement method based on key point detection

摘要

针对关键点检测的服装尺寸测量方法研究,提出基于 YOLO 改进的算法 RSV-YOLO,完成服装尺寸测量任务。结合 YOLOv9设计,实现双主干网络结构优化,增强特征点提取能力。提出 REMA 注意力机制利用残差学习思想,以捕获更加复杂的特征信息。RSPPF 特征融合,利用深度可分离卷积 DWConv和残差学习思想,降低复杂度,提升模型效率。由试验可知,RSV-YOLO 改进方案实现PointTop1值提高量为5.4%、PointTop2提高量为6.1%。改进算法 RSV-YOLOv10 的PointTop1值提高量为2.7%、PointTop2值提高量为2.6%,RSV-YOLOv11在PointTop1值上提高量为2.7%、PointTop2值提高量为2.7%。通过与注意力机制SE、CBAM、CA、EMA 对比可知,REMA 注意力机制在PointTop1值上提高量为2.7%、PointTop2值上提高量为3.3%,在关键点定位算法中获得最优表现。

Abstract

Aiming at the task of garment size measurement method for key point detection, we propose RSV-YOLO, an improved algorithm based on YOLO, to accomplish the task of garment size measurement. Drawing on the YOLOv9 design, this algorithm optimizes the dual backbone network structure and enhances the feature p oint extraction ability. We propose REMA attention mechanism utilizing residual learning ideas to capture more complex feature information. We employ RSPPF feature fusion, which integrates depthwise separable convolution (DWConv) and residual learning ideas. This design reduces complexity and improve model efficiency. As shown by the experiment, the RSV scheme combined with YOLOv8 achieves 5.4% improvement in PointTop1 value and 6.1% improvement in PointTop2 value. The improved algorithm RSV- YOLOv10 improves the PointTop1 value by 2.7% and PointTop2 by 2.6%, and RSV-YOLOv11 improves the PointTop1 value by 2.7% and PointTop2 by 2.7%. The comparison of the attention mechanisms SE, CBAM, CA, and EMA shows that the REMA attention mechanism improves 2.7% on the PointTop1 value and 3.3% on the PointTop2 value, and obtains the optimal performance in the key point localization algorithm.

关键词

服装关键点 / 服装尺寸 / REMA 注意力机制 / YOLOv11

Key words

garment keypoint / garment sizing / REMA / YOLOv11

参考文献

[1] 熊琳, 陈涛. 机器视觉技术在纺织品及服装研究中的应用[J]. 中国纤检, 2023(6): 110-112.

[2] XIONG L, CHEN T. Application of machine vision technology in textile and clothing research[J]. China Fiber Inspection, 2023(6): 110-112.

[3] XIA B, ZHANG Y L, WANG Y T, et al. Basic binary convolution unit for binarized image restoration network[EB/OL]. 2022: arXiv: 2210.00405. https://arxiv.org/abs/2210.00405.

[4] ZHAO X, XU B J, WU G X. Canny edge detection based on open CV[C]// 2017 13th IEEE International Conference on Electronic Measurement & Instruments (ICEMI). Yangzhou, China. IEEE, 2018: 53-56.

[5] HUSSAIN M. YOLO-v1 to YOLO-v8, the rise of YOLO and its complementary nature toward digital manufacturing and industrial defect detection[J]. Machines, 2023, 11(7): 677.

[6] WANG X L, SHRIVASTAVA A, GUPTA A. A-fast-RCNN: hard positive generation via adversary for object detection[C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA. IEEE, 2017: 3039-3048.

[7] LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot MultiBox detector[C]// Computer Vision-ECCV 2016. Cham: Springer, 2016: 21-37.

[8] OUYANG H D. DEYOv3: DETR with YOLO for real-time object detection[EB/OL]. 2023: arXiv: 2309. 11851. https://arxiv.org/abs/2309.11851.

[9] RONNEBERGER O, FISCHER P, BROX T. U-Net: convolutional networks for biomedical image segmentation[C]// Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015. Cham: Springer, 2015: 234-241.

[10] QIN X B, ZHANG Z C, HUANG C Y, et al. U2-Net: going deeper with nested U-structure for salient object detection[J]. Pattern Recognition, 2020, 106: 107404.

[11] SUN K, XIAO B, LIU D, et al. Deep high-resolution representation learning for human pose estimation[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach, CA, USA. IEEE, 2020: 5686-5696.

[12] QIAO S, WANG Y L, LI J. Real-time human gesture grading based on OpenPose[C]//2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI). Shanghai, China. IEEE, 2018: 1-6.

[13] 王政, 李飒, 陈洪界, 等. 基于改进DeeplabV3+ 的钢筋尺寸检测算法[J]. 地震工程与工程振动, 2023, 43(5): 240-248.

[14] WANG Z, LI S, CHEN H J, et al. Rebar size detection algorithm based on improved DeeplabV3+[J]. Earthquake Engineering and Engineering Dynamics, 2023, 43(5): 240-248.

[15] ZHANG Y, WU M L, LI J S, et al. Automatic non-destructive multiple lettuce traits prediction based on DeepLabV3+[J]. Journal of Food Measurement and Characterization, 2023, 17 (1): 636-652.

[16] 张海龙, 王念念. 多目标优化算法在服装尺寸调整与合体设计中的应用[J]. 染整技术, 2025, 47(7): 63-65.

[17] ZHANG H L, WANG N N. Application of multi-objective optimization algorithms in clothing size adjustment and fit design[J]. Textile Dyeing and Finishing Journal, 2025, 47 (7): 63-65.

[18] 王生伟, 张侃健. 基于角点检测的服装尺寸在线测量技术[J]. 信息技术与信息化, 2018(12): 73-75.

[19] WANG S W, ZHANG K J. Online measurement technology for garment size based on image processing[J]. Information Technology and Informatization, 2018(12): 73-75.

[20] 李鹏飞, 郑明智, 景军锋. 基于机器视觉的服装尺寸在线测量系统[J]. 毛纺科技, 2017, 45(3): 42-47.

[21] LI P F, ZHENG M Z, JING J F. Measurement system of garment dimension based on Machine Vision[J]. Wool Textile Journal, 2017, 45(3): 42-47.

[22] KRAUT S. Weiterer landwirtschaftspionier: Georg ferdinand forstner von dambenoy, Herr auf garnberg[J]. Württembergisch Franken, 2023, 104: 239-251.

[23] LOWE D G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2): 91-110.

[24] 冯采伶, 于施佳, 韩曙光. 基于深度学习的服装关键点实时检测模型[J]. 纺织学报, 2026, 47(1): 196-206.

[25] FENG C L, YU S J, HAN S G. Real-time detection model for clothing keypoints based on deep learning[J]. Journal of Textile Research, 2026, 47(1): 196-206.

[26] 陈润林, 史英杰, 杜方. YOLO-T-Shirt: 一种基于级联架构和融合几何信息的T 恤关键点检测方法[J]. 北京服装学院学报(自然科学版), 2024(2): 88-96.

[27] CHEN R L, SHI Y J, DU F. YOLO-T-shirt: a T-shirt landmark detection method based on cascade architecture and fusion geometry information[J]. Journal of Beijing Institute of Fashion Technology (Natural Science Edition), 2024(2): 88-96.

[28] 陶金之, 夏明, 王伟. 基于关键点检测的服装廓形识别[J]. 纺织学报, 2024, 45(6): 142-148.

[29] TAO J Z, XIA M, WANG W. Clothing silhouette recognition based on detection of key points[J]. Journal of Textile Research, 2024, 45(6): 142-148.

[30] WANG C Y, YEH I H, LIAO H M. YOLOv9: learning what you want to learn using programmable gradient information[EB/OL]. 2024: arXiv: 2402. 13616. https://arxiv.org/abs/2402.13616.

[31] WANG Z J, HE X W, LI Y, et al. EmbedFormer: embedded depth-wise convolution layer for token mixing[J]. Sensors, 2022, 22(24): 9854.

[32] GE Y Y, ZHANG R M, WANG X G, et al. DeepFashion2: a versatile benchmark for detection, pose estimation, segmentation and re-identification of clothing images[C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach, CA, USA. IEEE, 2020: 5332-5340.

[33] OUYANG D L, HE S, ZHANG G Z, et al. Efficient multi-scale attention module with cross-spatial learning[C]// ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Rhodes Island, Greece. IEEE, 2023: 1-5.

[34] DEWI C, JULI CHRISTANTO H. Combination of deep cross-stage partial network and spatial pyramid pooling for automatic hand detection[J]. Big Data and Cognitive Computing, 2022, 6(3): 85.

[35] 张庆, 林富生, 陈泽纯, 等. 基于轻量化 YOLOv8-GECS的织物疵点检测算法[J]. 东华大学学报(自然科学版), 2025, 51 (6): 54-61.

[36] ZHANG Q, LIN F S, CHEN Z C, et al. Fabric defect detection model based on lightweight YOLOv8-GECS[J]. Journal of Donghua University (Natural Science), 2025, 51 (6): 54-61.

[37] WANG A, CHEN H, LIU L H, et al. YOLOv10: real-time end-to-end object detection[EB/OL]. 2024: arXiv: 2405. 14458. https://arxiv.org/abs/2405.14458.

[38] KHANAM R, HUSSAIN M. YOLOv11: an overview of the key architectural enhancements [EB/OL]. (2024-10-23) [2025-01-07]. https://www.arxiv.org/abs/2410.17725.

[39] ZHANG Y, GUO Z Y, WU J Q, et al. Real-time vehicle detection based on improved YOLO v5[J]. Sustainability, 2022, 14(19): 12274.

[40] LI C Y, LI L L, JIANG H L, et al. YOLOv6: a single-stage object detection framework for industrial applications[EB/OL]. 2022: arXiv: 2209. 02976. https://arxiv.org/abs/2209.02976.

[41] WANG C Y, BOCHKOVSKIY A, LIAO H 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 (CVPR). Vancouver, BC, Canada. IEEE, 2023: 7464-7475.

[42] GE Z, LIU S T, WANG F, et al. YOLOX: exceeding YOLO series in 2021[EB/OL]. [2025-01-07]. 2021: arXiv: 2107. 08430. https://arxiv.org/abs/2107.08430.

[43] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA. IEEE, 2018: 7132-7141.

[44] 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 (CVPR). Nashville, TN, USA. IEEE, 2021: 13708-13717.

[45] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]// Computer Vision-ECCV 2018. Cham: Springer, 2018: 3-19.