• Laser & Optoelectronics Progress
  • Vol. 56, Issue 14, 141001 (2019)
Yongjie Ma*, Yunting Ma, and Jiahui Chen
Author Affiliations
  • College of Physics and Electronic Engineering, Northwest Normal University, Lanzhou, Gansu 730070, China
  • show less
    DOI: 10.3788/LOP56.141001 Cite this Article Set citation alerts
    Yongjie Ma, Yunting Ma, Jiahui Chen. Vehicle Recognition Based on Multi-Layer Features of Convolutional Neural Network and Support Vector Machine[J]. Laser & Optoelectronics Progress, 2019, 56(14): 141001 Copy Citation Text show less
    Structure of vehicle recognition method based on MCP-SVM hybrid model
    Fig. 1. Structure of vehicle recognition method based on MCP-SVM hybrid model
    Several images of samples. (a) Positive samples; (b) negative samples
    Fig. 2. Several images of samples. (a) Positive samples; (b) negative samples
    Comparison of accuracies and training loss curves
    Fig. 3. Comparison of accuracies and training loss curves
    NetworknameImageinputConvolutionkernelNetworklayerC1S1C2S2C3S3C4C5S5C6
    28×28383×32×22×22×23×3-2×22×22×2-
    48×48585×52×25×52×24×4-3×33×32×2-
    96×96585×52×25×52×25×52×25×55×52×2-
    28×28585×52×23×32×22×2-2×22×22×2-
    28×28787×72×22×22×22×2-2×22×22×2-
    28×28373×32×22×22×23×32×22×2---
    28×28393×32×22×2-2×22×22×23×3-2×2
    Table 1. Seven kinds of network structures based on AlexNet
    Classification network
    Training time /h5.811.5685.2106.37.5
    Accuracy rate /%97.8297.6294.0096.9296.7297.8797.76
    Table 2. Classification performance of seven kinds of networks
    LayerLayer inputConvolution kernalConvolution outputPoolingPooled output
    SizeNumStepSizeMode
    L1(C1+S1)28×28×33×396126×26×962×2Max13×13×96
    L2(C2+S2)13×13×962×2128112×12×1282×2Max6×6×128
    L3(C3+S3)6×6×1253×325614×4×2562×2Max2×2×256
    L4(C4)2×2×2562×225611×1×256---
    L5(Fc1)1×1×256------1024
    L6(Fc2)1024------1024
    L7(Softmax)1024------2
    Table 3. Structure of improved CNN model
    MethodTraining time /hAccuracy rate /%
    Using AlexNet model5196.92
    Using improved model6.397.87
    Table 4. Comparison and analysis of CNN models
    No.MethodAccuracyrate /%Testingtime /s
    1Method in Ref. [20]98.3288.36
    2MC-SVM98.72247.51
    3MCP-SVM98.7313.19
    Table 5. Comparison of classification performance of three methods
    MethodAccuracyrate /%Testingtime /s
    Improved CNN model97.87184
    Method in Ref. [21]91.751596
    Method in Ref. [4]92.331046
    Method in Ref. [6]94.72292
    MCP-SVM98.7313
    Table 6. Comparison of recognition rates of different methods in vehicle datasets
    Yongjie Ma, Yunting Ma, Jiahui Chen. Vehicle Recognition Based on Multi-Layer Features of Convolutional Neural Network and Support Vector Machine[J]. Laser & Optoelectronics Progress, 2019, 56(14): 141001
    Download Citation