• Laser & Optoelectronics Progress
  • Vol. 58, Issue 2, 0210015 (2021)
Jiao Yao* and Fengqin Yu
Author Affiliations
  • School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • show less
    DOI: 10.3788/LOP202158.0210015 Cite this Article Set citation alerts
    Jiao Yao, Fengqin Yu. Pedestrian Detection Based on Combination of Candidate Region Location and HOG-CLBP Features[J]. Laser & Optoelectronics Progress, 2021, 58(2): 0210015 Copy Citation Text show less
    Flow chart of proposed algorithm
    Fig. 1. Flow chart of proposed algorithm
    Candidate region windows generated by different methods. (a) Selective search strategy; (b) restriction of aspect ratio
    Fig. 2. Candidate region windows generated by different methods. (a) Selective search strategy; (b) restriction of aspect ratio
    Recognition rate and time of PCA-HOG in different dimensions. (a) Identification rate; (b) time
    Fig. 3. Recognition rate and time of PCA-HOG in different dimensions. (a) Identification rate; (b) time
    Recognition rate and time of PCA-CLBP in different dimensions. (a) Identification rate; (b) time
    Fig. 4. Recognition rate and time of PCA-CLBP in different dimensions. (a) Identification rate; (b) time
    Comparison of DET curves of different algorithms
    Fig. 5. Comparison of DET curves of different algorithms
    Detection results of proposed algorithm on some images
    Fig. 6. Detection results of proposed algorithm on some images
    PCA-HOG dimension203040506080100200300
    Recognition rate0.89000.92070.92080.91990.92110.92270.91860.91830.9181
    Time /s0.12900.14000.16700.35300.56400.76601.50702.03304.3040
    Table 1. Recognition rate and recognition speed of PCA-HOG in different dimensions
    PCA-CLBP dimension123457102030
    Recognition rate0.76290.77940.79640.80230.80230.80290.79510.79500.7948
    Time /s0.11030.11570.12090.12520.14670.19550.25340.56840.9430
    Table 2. Recognition rate and recognition speed of PCA-CLBP in different dimensions
    AlgorithmRecognition rate /%Test time per image /s
    Ref.[2]89.008.560
    Ref.[6]94.703.500
    Ref.[10]90.100.008
    Ref.[11]75.340.460
    Ref.[22]93.507.200
    Ref.[23]93.603.870
    Proposed algorithm96.701.960
    Table 3. Recognition rate and recognition speed of different algorithms
    Jiao Yao, Fengqin Yu. Pedestrian Detection Based on Combination of Candidate Region Location and HOG-CLBP Features[J]. Laser & Optoelectronics Progress, 2021, 58(2): 0210015
    Download Citation