Transfer learning with pre-trained deep convolutional neural networks for the automatic assessment of liver steatosis in ultrasound images.

Journal: Medical ultrasonography
Published Date:

Abstract

AIM: In this paper we proposed different architectures of convolutional neural network (CNN) to classify fatty liver disease in images using only pixels and diagnosis labels as input. We trained and validated our models using a dataset of 629 images consisting of 2 types of liver images, normal and liver steatosis.

Authors

  • Elena Codruta Constantinescu
    Research Center of Gastroenterology and Hepatology, University of Medicine and Pharmacy Craiova, Romania. constantinescu.codruta@yahoo.com.
  • Anca-Loredana Udriștoiu
    Faculty of Automation, Computers and Electronics, University of Craiova, Craiova, Romania. anca_soimu@yahoo.com.
  • Ștefan Cristinel Udriștoiu
    Faculty of Automation, Computers and Electronics, University of Craiova, Craiova, Romania. stefan@software.ucv.ro.
  • Andreea Valentina Iacob
    Faculty of Automation, Computers and Electronics, University of Craiova, Craiova, Romania. andr33a_soimu@yahoo.com.
  • Lucian Gheorghe Gruionu
    Faculty of Mechanics, University of Craiova, Craiova, Romania. lgruionu@yahoo.com.
  • Gabriel Gruionu
    Faculty of Mechanics, University of Craiova, Craiova, Romania. gruionu@gmail.com.
  • Larisa Săndulescu
    Research Center of Gastroenterology and Hepatology, University of Medicine and Pharmacy Craiova, Romania. larisasandulescu@yahoo.com.
  • Adrian Săftoiu
    Research Center of Gastroenterology and Hepatology, University of Medicine and Pharmacy Craiova, Romania. adriansaftoiu@gmail.com.