Human Action Recognition Based on Global and Local Features

Liu Fan; Yu Fengqin

doi:10.3788/LOP57.021004

Journals >Laser & Optoelectronics Progress >Volume 57 >Issue 2 >Page 21004 > Article

Laser & Optoelectronics Progress
Vol. 57, Issue 2, 21004 (2020)

Human Action Recognition Based on Global and Local Features

Liu Fan^* and Yu Fengqin

Author Affiliations

School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China

show less

DOI: 10.3788/LOP57.021004 Cite this Article Set citation alerts

Liu Fan, Yu Fengqin. Human Action Recognition Based on Global and Local Features[J]. Laser & Optoelectronics Progress, 2020, 57(2): 21004 Copy Citation Text

show less

Fig. 1. Flow chart of algorithm

Download full size

Fig. 2. SURF extraction

Download full size

Fig. 3. Comparison of HOG features before and after improvement

Download full size

Fig. 4. Recognition accuracies of proposed algorithm in different datasets. (a) Confusion matrix on KTH dataset; (b) confusion matrix on UCF Sports dataset; (c) confusion matrix on SBU Kinect Interaction dataset

Download full size

Method	η on KTH /%	Method	η on UCF sports /%	Method	η on SBU Kinect Interaction /%
Method in Ref. [16]	95.6	Ref. [16]	88.5	Ref. [15]	80.3
Method in Ref. [17]	95.6	Ref. [18]	92.7	Ref. [22]	89.4
Method in Ref. [18]Method in Ref. [19]Proposed method	96.897.896.7	Ref. [20]Ref. [21]Proposed	95.897.594.4	Ref. [23]Ref. [24]Proposed	94.898.590.8

Table 1. Recognition accuracies of different methods

Liu Fan, Yu Fengqin. Human Action Recognition Based on Global and Local Features[J]. Laser & Optoelectronics Progress, 2020, 57(2): 21004

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information