Author Affiliations
1 Key Laboratory of Airborne Optical Imaging and Measurement, Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin 130033, China2 University of Chinese Academy of Sciences, Beijing 100049, Chinashow less
Fig. 1. Schematic of linear inseparable problem
Fig. 2. Schematic of original space to high-dimensional feature space
Fig. 3. Schematic of near linear separable problem
Fig. 4. Feature extraction diagram. Results of (a) gray processing, (b) Gamma correction, and (c) gradient image
Fig. 5. Graph of kernel function. (a) Polynomial kernel function curve; (b) Gaussian kernel function curve
Fig. 6. Graph of combined kernel function
Fig. 7. Comparison of recognition rate
Fig. 8. Line chart of the change of recognition rate with C. (a) Polynomial kernel function; (b) Gaussian kernel function; (c) combined kernel function
Fig. 9. Comparison of the recognition rate between proposed algorithm and traditional algorithms
Fig. 10. Results of people detection
d | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
---|
Recognition rate /% | 97.50 | 96.73 | 95.99 | 95.20 | 94.43 | 93.60 | 92.80 | 92.00 | 91.54 |
|
Table 1. Change of recognition rate with d
σ | 0.1 | 0.2 | 0.5 | 0.7 | 1.0 | 1.5 | 2.0 | 3.0 | 4.0 |
---|
Recognition rate /% | 96.63 | 97.20 | 96.84 | 96.00 | 95.72 | 95.40 | 95.00 | 94.64 | 94.12 |
|
Table 2. Change of recognition rate with σ
α1 | 0.1 | 0.2 | 0.3 | 0.4 | 0.5 | 0.6 | 0.7 | 0.8 | 0.9 |
---|
Recognition rate /% | 96.24 | 96.76 | 97.34 | 97.22 | 97.84 | 97.95 | 97.92 | 98.25 | 98.50 |
|
Table 3. Change of recognition rate with α1