基于小目标增强金字塔的高精度定位手部姿态估计
1. 四川师范大学, 计算机科学学院, 四川成都 610101
2. 四川师范大学, 物理与电子工程学院, 四川成都 610101
| 摘 要： | 手部关节姿态评估是手部关节活动度司法鉴定中的关键环节，在司法鉴定领域具有重要意义。为克服传统方法对专业医师高度依赖的局限，本文提出了一种基于计算机视觉的改进型手部姿态评估方法。该方法聚焦于手部21个关键点，并针对其在图像中尺寸较小、易被遮挡、特征信息有限等问题，构建了优化的检测框架。具体而言，我们设计了小目标增强特征金字塔，以缓解小关键点特征易丢失的难题；引入SPDConv模块以保留细节特征；并结合Omni-FSAK（Omni-Frequency-Spatial Attention Kernel）模块，实现高效的多尺度特征融合，从而显著提升对模糊、遮挡及细小关键点的感知能力。在检测头设计方面，除分类分支、边界框回归分支和关键点预测分支外，我们进一步引入了位置质量校准器（LQC, Localization Quality Calibrator），充分利用分类分支输出与回归分布信息，对预测结果的可靠性进行评估。该机制有效弥补了分类置信度无法准确反映位置质量的不足，使模型在相同网络规模下能够更稳健地输出高质量的检测与关键点预测结果。实验在FreiHAND和KeypointsHand数据集上验证了所提出方法的有效性，相较基线方法，准确率提升了3.15%，召回率提升了1.74%，充分验证了改进方案的可行性。 |
| 关 键 词： | 手部关键点检测; 小目标检测; 司法鉴定; 姿态估计; 定位质量估计 |
| DOI： | 10.57237/j.cst.2026.01.001 |
View:
Downloads:
收稿日期: 2025-11-24
接受日期: 2025-12-26
在线出版日期: 2026-01-14
High-Precision Hand Pose Estimation Based on Small Object Augmented Feature Pyramid
1. School of Computer Science, Sichuan Normal University, Chengdu 610101, China
2. School of Physics and Electronic Engineering, Sichuan Normal University, Chengdu 610101, China
| Abstract: | Hand joint pose assessment is a critical step in the forensic identification of range of motion, holding significant importance in the field of forensic science. To overcome the limitations of traditional methods that rely heavily on professional physicians, this paper proposes an improved hand pose assessment method based on computer vision. Focusing on 21 hand keypoints, the method addresses challenges such as their small size in images, susceptibility to occlusion, and limited feature information by constructing an optimized detection framework. Specifically, a Small Object Augmented Feature Pyramid is designed to alleviate the difficulty of losing small keypoint features; the SPDConv module is introduced to preserve detailed features; and the Omni-Frequency-Spatial Attention Kernel (Omni-FSAK) module is incorporated to achieve efficient multi-scale feature fusion, thereby significantly enhancing the perception capability for blurred, occluded, and tiny keypoints. In terms of detection head design, in addition to the classification branch, bounding box regression branch, and keypoint prediction branch, we further introduce a Localization Quality Calibrator (LQC). This mechanism fully utilizes the outputs from the classification branch and regression distribution information to evaluate the reliability of prediction results. It effectively compensates for the insufficiency of classification confidence in accurately reflecting localization quality, enabling the model to output high-quality detection and keypoint prediction results more robustly under the same network scale. Experiments on the FreiHAND and KeypointsHand datasets verify the effectiveness of the proposed method. Compared with the baseline method, the accuracy is improved by 3.15% and the recall rate is increased by 1.74%, fully validating the feasibility of the improved scheme. |
| Keywords: | Hand Keypoint Detection; Small Object Detection; Forensic Identification; Pose Estimation; Localization Quality Estimation |
基金
| 1. | 国家社会科学基金一般项目 (20BMZ092) |
参考文献
| [1] | Kadam P, Fang G, Amirabdollahian F, et al. Hand Pose Detection Using YOLOv8-pose [C] // 2024 IEEE Conference on Engineering Informatics (ICEI). IEEE, 2024: 1-6. |
| [2] | Che W, Zhang H, Wu B, et al. SFG-YOLOv8: efficient and lightweight small-feature gesture keypoint detector [J]. Journal of King Saud University Computer and Information Sciences, 2025, 37(4): 44. |
| [3] | Myanganbayar B, Mata C, Dekel G, et al. Partially occluded hands: A challenging new dataset for single-image hand pose estimation [C] // Asian Conference on Computer Vision. Cham: Springer International Publishing, 2018: 85-98. |
| [4] | Jia X, Feng J, Liang B. Hand keypoint detection with super resolution [C] // Proceedings of the 7th International Conference on Cyber Security and Information Engineering. 2022: 398-401. |
| [5] | Romero J, Tzionas D, Black M J. Embodied hands: Modeling and capturing hands and bodies together [J]. arXiv preprint arXiv: 2201.02610, 2022. |
| [6] | JING L, WANG B. EMNet: Edge-guided multi-level network for salient object detection in low-light images [J]. Image and Vision Computing, 2024, 143: 104933. |
| [7] | 侯贝贝.基于手部姿态估计的手关节角度测量系统研究[D].中国科学院大学 (中国科学院西安光学精密机械研究所), 2024. https://doi.org/10.27605/d.cnki.gkxgs.2024.000060 |
| [8] | Chen M, Shuang F, Li S, et al. Robust dexterous hand control strategy cascading bare hand pose estimation and joint jitter suppression [J]. Robotics and Autonomous Systems, 2025: 105189. |
| [9] | 任建华,曹佳惠,贾迪.基于掩码提示和注意力的手部姿态估计[J/OL].计算机应用, 1-11 [2025-09-29]. https://doi.org/10.11772/j.issn.1001-9081.2024111715 |
| [10] | 郑泉石,金城.基于自适应预测的2D人体姿态估计[J].计算机科学, 2023, 50(S2): 162-168. https://doi.org/CNKI:SUN:JSJA.0.2023-S2-023 |
| [11] | 刘佳, 石豪, 陈大鹏,等. 基于YOLOv3-HM的手部姿态估计方法研究 [J]. 测控技术, 2023(004): 60-66, 87. |
| [12] | 倪广兴, 徐华, 王超. 融合改进YOLOv5及Mediapipe的手势识别研究 [J]. 计算机工程与应用, 2024, 60(07): 108-118. |
| [13] | Ding J, Niu S, Nie Z, et al. Research on human posture estimation algorithm based on YOLO-Pose [J]. Sensors, 2024, 24(10): 3036. |
| [14] | Zhu X, Lyu S, Wang X, et al. TPH-YOLOv5: Improved YOLOv5 based on transformer prediction head for object detection on drone-captured scenarios [C] // Proceedings of the IEEE/CVF international conference on computer vision. 2021: 2778-2788. |
| [15] | Guan X, Shen H, Nyatega C O, et al. Repeated Cross-Scale Structure-Induced Feature Fusion Network for 2D Hand Pose Estimation [J]. Entropy, 2023, 25(5): 724. |
| [16] | Zimmermann C, Ceylan D, Yang J, et al. Freihand: A dataset for markerless capture of hand pose and shape from single rgb images [C] // Proceedings of the IEEE/CVF international conference on computer vision. 2019: 813-822. |
| [17] | Zhang Y, Du S, He H. Rotator-YOLOv5: Improved YOLOv5 for Vehicle and Vessel Detection in UAV Images [C] // 2024 Fourth International Conference on Digital Data Processing (DDP). IEEE, 2024: 156-161. |
| [18] | Zhao P. SPFFNet: Strip perception and feature fusion spatial pyramid pooling for fabric defect detection [J]. arXiv preprint arXiv: 2502.01445, 2025. |
| [19] | Yang G, Lei J, Zhu Z, et al. AFPN: Asymptotic feature pyramid network for object detection [C] // 2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC). IEEE, 2023: 2184-2189. |
| [20] | 李少东, 罗凯, 黄远智, 等. 复杂交互场景下融合关节遮挡信息的手部姿态估计研究 [J]. 计算机学报, 2025, 48(05): 1212-1231. |
