Application of Transfer Learning for Object Recognition Using Convolutional Neural Networks

  • Jesús Alfonso López Sotelo
  • Nicolás Díaz Salazar
  • Gustavo Andres Salazar GomezEmail author
Conference paper
Part of the Communications in Computer and Information Science book series (CCIS, volume 833)


In this work, the transfer learning technique is used to create a computational tool that recognizes the objects of the automation laboratory of the Universidad Autónoma de Occidente in real time. As a pre-trained neural net, the Inception-V3 is used as a feature extractor in the images and on the other hand a softmax classifier is trained, this contains the classes that are going to be recognized. It was used Tensorflow platform with gpu in Python natively in Windows 10 and Opencv library for the use of video camera and other tools.


Transfer learning Softmax Inception-V3 Tensorflow Neural networks Convolutional neural network 


  1. 1.
    Castrillon, W., Alvarez, D., López, A.: Técnicas de extracción de características en imágenes para el reconocimiento de expresiones faciales. In: Scientia Et Technica, vol. XIV, no. 38, pp. 7–12 (2008)Google Scholar
  2. 2.
    Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., Torralba, A.: Object detectors emerge in deep scene CNNs (2015)Google Scholar
  3. 3.
    Brown, L.: Deep learning with GPUs (2015)Google Scholar
  4. 4.
    Loncomilla, P.: Deep learning: redes convolucionales (2016)Google Scholar
  5. 5.
    The data science blog. An intuitive explanation of convolutional neural networks.
  6. 6.
    Szegedy, C., et al.: Going deeper with convolutions (2014)Google Scholar
  7. 7.
    Alemi, A.: Improving inception and image classification in tensorflow (2016)Google Scholar
  8. 8.
    Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., Wojna, Z.: Rethinking the inception architecture for computer vision (2015)Google Scholar
  9. 9.
    Tensorflow how to retrain inception´s final layer for new categories.

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Jesús Alfonso López Sotelo
    • 1
  • Nicolás Díaz Salazar
    • 1
  • Gustavo Andres Salazar Gomez
    • 1
    Email author
  1. 1.Automation and Electronics DepartmentUniversidad Autónoma De OccidenteCaliColombia

Personalised recommendations