3D gesture trajectory recognition based on point context descriptor

被引：0

作者：

Mao X. ^{[1
]}

Li C. ^{[1
]}

Wu X. ^{[1
]}

机构：

[1] School of Electronic and Information Engineering, Beijing University of Aeronautics and Astronautics, Beijing

来源：

Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition) | 2016年 / 44卷 / 08期

关键词：

3D gesture trajectory recognition; Curvature window; Feature vector; Pattern recognition; Point context descriptor;

D O I：

10.13245/j.hust.160811

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

A method for 3D (three dimension) gesture trajectory recognition based on point context descriptor of curvature window was proposed. By using curvature window to choose the context points of a gesture trajectory, the method can not only reduce the dimension of the descriptor but also raise the recognition rate of 3D gesture trajectory. Firstly, the U-chord curvature of all points on a 3D gesture trajectory were calculated. Then, points which have the maximum curvature value and separate to each other were chosen as the context points of a gesture trajectory. Lastly, the Euclidean distances between every point on the trajectory and the context points were regarded as the feature vector of the point. The shape feature of a 3D gesture trajectory was extracted, and the support vector machine (SVM) was applied to classification and recognition. Experimental results show that the average recognition rate can reach 92.83% when the method is applied to 8 kinds of 3D gesture trajectories chosen randomly from the Australian sign language (ASL) data set. Besides, the recognition result is invariant to rotation, scaling and translation (RST) transformation. © 2016, Editorial Board of Journal of Huazhong University of Science and Technology. All right reserved.

引用

页码：52 / 57

页数：5

共 8 条

[1]

Akhter I., Sheikh Y., Khan S., Et al., Trajectory space: a dual representation for nonrigid structure from motion, IEEE Transactions on Pattern Analysis and Machine Intelligence, 33, 7, pp. 1442-1456, (2011)

[2]

Harding P.R., Ellis T.J., Recognizing hand gesture using Fourier descriptors, Proceedings of IEEE Conference on Pattern Recognition, 3, pp. 286-289, (2004)

[3]

Hung K.C., The generalized uniqueness wavelet descriptor for planar closed curves, IEEE Transactions on Image Processing, 9, 5, pp. 834-845, (2000)

[4]

Bashir F.I., Khokhar A.A., Schonfeld D., View-invariant motion trajectory-based activity classification and recognition, Ltimedia Systems, 12, 1, pp. 45-54, (2006)

[5]

Wu X.Y., Mao X., Chen L.J., Et al., Point context: an effective shape descriptor for RST-invariant trajectory recognition, Journal of Mathematical Imaging and Vision, 55, 3, pp. 1-14, (2016)

[6]

Wu S., Li Y.F., Flexible signature descriptions for adaptive motion trajectory representation, perception and recognition, Pattern Recognition, 42, 1, pp. 194-214, (2009)

[7]

Yang J., Li Y., Wang K., Et al., Mixed signature: an invariant descriptor for 3D motion trajectory perception and recognition, Proceedings of IEEE Conference on Industrial Engineering and Engineering Management, pp. 1952-1956, (2010)

[8]

Lin W.Y., Hsieh C.Y., Kernel-based representation for 2D/3D motion trajectory retrieval and classification, Pattern Recognition, 46, 3, pp. 662-670, (2013)

← 1 →