[1]苏 琳,邹 静,许 媚.VR中基于改进SIFT特征匹配的手势检测与跟踪[J].南京师大学报(自然科学版),2025,48(02):83-90.[doi:10.3969/j.issn.1001-4616.2025.02.009]
　Su Lin,Zou Jing,Xu Mei.Gesture Detection and Tracking Based on Improved SIFT Feature Matching in VR[J].Journal of Nanjing Normal University(Natural Science Edition),2025,48(02):83-90.[doi:10.3969/j.issn.1001-4616.2025.02.009]
点击复制

VR中基于改进SIFT特征匹配的手势检测与跟踪()

《南京师大学报（自然科学版）》[ISSN:1001-4616/CN:32-1239/N]

卷:
48
期数:
2025年02期
页码:
83-90
栏目:
计算机科学与技术
出版日期:
2025-04-15

文章信息/Info

Title:
Gesture Detection and Tracking Based on Improved SIFT Feature Matching in VR
文章编号:
1001-4616(2025)02-0083-08
作者:
苏 琳1; 邹 静2; 许 媚3
(1.闽南科技学院艺术设计学院,福建 泉州 362332)
(2.贵州大学计算机科学与技术学院,贵阳 贵州 550025)
(3.西北师范大学教育科学学院,甘肃 兰州 730070)
Author(s):
Su Lin1; Zou Jing2; Xu Mei3
(1.School of Art and Design,Minnan University of Science and Technology,Quanzhou 362332,China)
(2.School of Computer Science and Technology,Guizhou University,Guiyang 550025,China)
(3.College of Education Sciences,Northwest Normal University,Lanzhou 730070,China)
关键词:
虚拟现实; 手势检测; 改进SIFT算法; 特征匹配; 人机交互
Keywords:
virtual reality; gesture detection; improve SIFT algorithm; feature matching; human-computer interaction
分类号:
TP391.9
DOI:
10.3969/j.issn.1001-4616.2025.02.009
文献标志码:
A
摘要:
随着虚拟现实(virtual reality,VR)技术的迅猛发展,对自然人机交互的需求日益增长,其中手势识别技术扮演着至关重要的角色. 它不仅要求高度的准确性,还必须保证实时响应,以确保用户能够获得流畅的交互体验. 本研究提出了一种创新的手势检测与跟踪方法,该方法基于改进的尺度不变特征变换(scale invariant feature transform,SIFT)特征匹配技术,专门针对VR环境中的手势识别进行了优化. 首先,本文对SIFT算法进行了深入的改进,通过引入先进的描述子来增强特征的区分度,这使得算法能够更准确地捕捉到手势的关键特征. 然后,为了进一步提升匹配的准确性,我们精心设计了特征匹配策略,优化了特征点之间的对应关系,确保了在复杂场景下也能实现高效匹配. 最后,针对实时性的需求,本文开发了一套算法优化策略,通过调整算法流程和计算方式,确保了算法即使在动态和多变的环境中也能保持高效稳定的运行,从而满足了实时手势跟踪的应用需求. 实验结果表明,所提模型的预测准确率为0.926,表现出了优异的预测性能.
Abstract:
With the rapid development of virtual reality(VR)technology,the demand for natural human-computer interaction is increasing,and gesture recognition technology plays a crucial role. It not only requires high accuracy,but also must ensure real-time response to ensure that users can have a smooth interactive experience. This study proposes an innovative gesture detection and tracking method based on an improved Scale Invariant Feature Transform(SIFT)feature matching technique,specifically optimized for gesture recognition in VR environments. Firstly,this article has made in-depth improvements to the SIFT algorithm by introducing advanced descriptors to enhance feature discrimination,which enables the algorithm to more accurately capture key features of gestures. Then,in order to further improve the accuracy of matching,we carefully designed a feature matching strategy,optimized the correspondence between feature points,and ensured efficient matching even in complex scenes. Finally,in response to the real-time requirements,this article developed an algorithm optimization strategy that ensures efficient and stable operation of the algorithm even in dynamic and changing environments by adjusting the algorithm flow and calculation methods,thus meeting the application needs of real-time gesture tracking. The experimental results show that the prediction accuracy of the proposed model is 0.926,demonstrating excellent predictive performance.

参考文献/References:

[1]高方远,何立火. 基于深度网络和视觉特性的无参考图像质量评价方法[J]. 南京师大学报(自然科学版),2019,42(3):20-26.
[2]龚新儒,张德育,谢少威. 虚拟现实中基于轻量级神经网络的手势识别[J]. 南京理工大学学报(自然科学版),2024,48(3):367-373.
[3]KUNZ A,ALAVI A,SINN P. Integrating pointing gesture detection for enhancing brainstorming meetings using kinect and pixelsense[J]. Procedia cirp,2014,25:205-212.
[4]PARADA R,MELIA S J. Gesture detection using passive rfid tags to enable people-centric IoT applications[J]. IEEE communications magazine,2017,55(2):56-61.
[5]KIM K,KIM J,CHOI J,et al. Depth camera-based 3D hand gesture controls with immersive tactile feedback for natural mid-air gesture interactions[J]. Sensors,2015,15(1):1022-1034.
[6]SONG W,LU Z,LI J,et al. Hand gesture detection and tracking methods based on background subtraction[J]. Lecture notes in electrical engineering,2014,309:485-490.
[7]CHEN Q,GEORGANAS N D,PETRIU E M. Hand gesture recognition using haar-like features and a stochastic context-free grammar[J]. Instrumentation and measurement,2008,57(8):1562-1571.
[8]SHERLY L A,JAYA T. An efficient indoor scene character recognition using Bayesian interactive search algorithm-based adaboost-CNN classifier[J]. Neural computing and applications,2021,33(22):15345-56.
[9]WANG J,JIA Z,LAI H,et al. A real time target face tracking algorithm based on saliency detection and Camshift[J]. Multimedia tools and applications,2023,82(28):43599-624.
[10]FANG B. Vehicle rear end collision warning method based on MeanShift and Kalman filter tracking[J]. International journal of vehicle design,2024,95(1):1-21.
[11]ALI A,GHOSH A,CHAUDHURI S S. Real-time tracking of moving objects through efficient scale space adaptation and normalized correlation filtering[J]. Signal,image and video processing,2024,18(1):679-89.
[12]邬忠萍,刘新厂,郝宗波. 基于并行CNN和识别策略优化的车牌识别方法研究[J]. 南京师大学报(自然科学版),2023,46(3):98-104.
[13]鲁东生,龙华. 一种基于谱图SIFT的同源频谱监测数据判定方法[J]. 计算机科学,2024,51(S1):777-783.
[14]PIRAN Z,NITZAN M. SiFT:uncovering hidden biological processes by probabilistic filtering of single-cell data[J]. Nature communications,2024,15(1):760-772.
[15]ZHANG Y,ZHU D,WANG Y L. Improved fast feature matching method of SIFT[J]. Control and automation pubilcation group,2008,24(11):220-222.
[16]王锟,宋永红,郑斐,等. 基于卷积神经网络的嵌入式手势检测算法[J]. 计算机工程与应用,2019,55(4):137-141,178.
[17]姬晓飞,张旭,李俊鹏. 基于多域卷积神经网络跟踪的动态手势识别[J]. 沈阳航空航天大学学报,2021,38(5):51-57.
[18]陈威. 一种基于光流的动态手势识别方法[J]. 哈尔滨师范大学自然科学学报,2015,31(6):81-86.

相似文献/References:

[1]刘晓艳,闾国年,张宏,等.南京师范大学虚拟新校区设计与建设[J].南京师大学报(自然科学版),2001,24(01):114.
　Liu Xiaoyan,Lu Guonian,Zhang Hong,et al.Design and Construction of Virtual Xianlin Campus of Nanjing Normal University[J].Journal of Nanjing Normal University(Natural Science Edition),2001,24(02):114.

备注/Memo

备注/Memo:
收稿日期:2024-11-26.
基金项目:福建省自然科学基金面上项目(2023J011406).
通讯作者:苏琳,讲师,研究方向:虚拟现实,人工智能. E-mail:3051700287@qq.com
更新日期/Last Update: 2025-04-15