Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Laser & Optoelectronics Progress, Volume. 56, Issue 23, 231502(2019)

Object Classification Based on Multitask Convolutional Neural Network

Miaohui Zhang1,2, Bo Zhang1、*, and Chengcheng Gao1
Author Affiliations
  • 1Henan Key Laboratory of Big Data Analysis and Processing, Henan University, Kaifeng, Henan 475004, China
  • 2Institute of Data and Knowledge Engineering, Henan University, Kaifeng, Henan 475004, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (10)
    Equations (0)
    References (19)
    Cited By (2)
    Tools
    Paper Information
    Topics
    Figures & Tables(10)
    Structure of MTCNN

    Fig. 1. Structure of MTCNN

    Download full size
    Detection of image detail areas. (a) Original image; (b) saliency detection; (c) corner detection

    Fig. 2. Detection of image detail areas. (a) Original image; (b) saliency detection; (c) corner detection

    Download full size
    Fine-grained images

    Fig. 3. Fine-grained images

    Download full size
    Accuracy changes of car-type and car-color datasets under PReLu. (a) Accuracy changes of car-type; (b) accuracy changes of car-color

    Fig. 4. Accuracy changes of car-type and car-color datasets under PReLu. (a) Accuracy changes of car-type; (b) accuracy changes of car-color

    Download full size
    Car Dataset

    Fig. 5. Car Dataset

    Download full size
    Accuracy changes of MTCNN training for Car Dataset

    Fig. 6. Accuracy changes of MTCNN training for Car Dataset

    Download full size
    Accuracy changes of MTCNN* training for Car Dataset

    Fig. 7. Accuracy changes of MTCNN* training for Car Dataset

    Download full size
    Multi-label outputs of samples

    Fig. 8. Multi-label outputs of samples

    Download full size

    • Table 1. Group experiments of Car Dataset

      View table

      Table 1. Group experiments of Car Dataset

      DatasetModelClassTrainTestAccuracy /%
      Car-multi(type)MTCNN*2159969790.23
      Car-multi(color)MTCNN*5150069380.75
      Car-multi(type)MTCNN2159969781.03
      Car-multi(color)MTCNN5150069376.18
      Car-typeAlexNet2159969777.75
      Car-colorAlexNet5150069375.06

    • Table 2. Performance of Car Dataset under different networks

      View table

      Table 2. Performance of Car Dataset under different networks

      DatasetModelNet-DeepBatchsModel size /MbitAccuracy /%
      Car-multi(type)MTCNN*8200021790.23
      Car-multi(color)MTCNN*8200021780.75
      Car-typeAlexNet8150020277.75
      Car-colorAlexNet8150020275.06
      Car-typeGoogleNet-v12225004890.15
      Car-colorGoogleNet-v12225004878.78
      Car-typeVGG-1619300052788.56
      Car-colorVGG-1619300052777.17
    Tools

    Get Citation

    Copy Citation Text

    Miaohui Zhang, Bo Zhang, Chengcheng Gao. Object Classification Based on Multitask Convolutional Neural Network[J]. Laser & Optoelectronics Progress, 2019, 56(23): 231502

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Set citation alerts for article
    Save article for my favorites
    Paper Information

    Category: Machine Vision

    Received: Apr. 12, 2019

    Accepted: May. 27, 2019

    Published Online: Nov. 27, 2019

    The Author Email: Bo Zhang (zhangbo208@163.com)

    DOI:10.3788/LOP56.231502

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology