The Utility of Applying Various Image Preprocessing Strategies to Reduce the Ambiguity in Deep Learning-based Clinical Image Diagnosis.

Journal: Magnetic resonance in medical sciences : MRMS : an official journal of Japan Society of Magnetic Resonance in Medicine
Published Date: May 10, 2019

Abstract

PURPOSE: A general problem of machine-learning algorithms based on the convolutional neural network (CNN) technique is that the reason for the output judgement is unclear. The purpose of this study was to introduce a strategy that may facilitate better understanding of how and why a specific judgement was made by the algorithm. The strategy is to preprocess the input image data in different ways to highlight the most important aspects of the images for reaching the output judgement.

Authors

  • Yasuhiko Tachibana
    National Institute of Radiological Sciences, Chiba, Chiba, Japan.
  • Takayuki Obata
    Applied MRI Research, Department of Molecular Imaging and Theranostics, National Institute of Radiological Sciences, National Institutes for Quantum and Radiological Science and Technology.
  • Jeff Kershaw
    Applied MRI Research, Department of Molecular Imaging and Theranostics, National Institute of Radiological Sciences, National Institutes for Quantum and Radiological Science and Technology.
  • Hironao Sakaki
    Kansai Photon Science Institute, National Institutes for Quantum and Radiological Science and Technology.
  • Takuya Urushihata
    Department of Functional Brain Imaging Research, National Institute of Radiological Sciences, National Institutes for Quantum and Radiological Science and Technology.
  • Tokuhiko Omatsu
    Applied MRI Research, Department of Molecular Imaging and Theranostics, National Institute of Radiological Sciences, National Institutes for Quantum and Radiological Science and Technology.
  • Riwa Kishimoto
    Applied MRI Research, Department of Molecular Imaging and Theranostics, National Institute of Radiological Sciences, National Institutes for Quantum and Radiological Science and Technology.
  • Tatsuya Higashi
    Faculty of Pharmaceutical Sciences, Tokyo University of Science, 2641 Yamazaki, Noda-shi, Chiba 278-8510, Japan. Electronic address: higashi@rs.tus.ac.jp.

Keywords

External Resources

Popular Topics
Recent Journals