A Preliminary Study on Deep Transfer Learning Applied to Image Classification for Small Datasets

  1. Miguel Ángel Molina 1
  2. Gualberto Asencio Cortés 1
  3. José Cristobal Riquelme Santos 2
  4. Francisco Martínez-Álvarez 1
  1. 1 Universidad Pablo de Olavide
    info

    Universidad Pablo de Olavide

    Sevilla, España

    ROR https://ror.org/02z749649

  2. 2 Universidad de Sevilla
    info

    Universidad de Sevilla

    Sevilla, España

    ROR https://ror.org/03yxnpp24

Book:
15th International Conference on Soft Computing Models in Industrial and Environmental Applications (SOCO 2020): Burgos, Spain ; September 2020
  1. Álvaro Herrero Cosío (ed. lit.)
  2. Carlos Cambra Baseca (ed. lit.)
  3. Daniel Urda Muñoz (ed. lit.)
  4. Javier Sedano Franco (ed. lit.)
  5. Héctor Quintián Pardo (ed. lit.)
  6. Emilio Santiago Corchado Rodríguez (ed. lit.)

Publisher: Springer Suiza

ISBN: 978-3-030-57801-5

Year of publication: 2021

Pages: 741-750

Congress: International Conference on Soft Computing Models in Industrial and Environmental Applications SOCO (15. 2020. Burgos)

Type: Conference paper

DIALNET GOOGLE SCHOLAR

Abstract

A new transfer learning strategy is proposed for image classification in this work, based on an 8-layer convolutional neural network. The transfer learning process consists in a training phase of the neural network on a source dataset of images. Then, the last two layers are retrained using a different small target dataset of images. A preliminary study was conducted to train and test the transfer learning proposal on Malaria cell images for a binary classification problem. The methodology proposed has provided a 6.76% of improvement with respect to other three different strategies of training non-transfer learning models. The results achieved are quite promising and encourage to conduct further research in this field.

Data source: Dialnet