Breast Ultrasound Region of Interest Detection and Lesion Localisation

Authors Organisations
  • Moi Hoon Yap(Author)
    Manchester Metropolitan University
  • Manu Goyal(Author)
    Manchester Metropolitan University
  • Fatima Osman(Author)
    Sudan University of Science and Technology
  • Robert Marti(Author)
    University of Girona
  • Erika R. E. Denton(Author)
    Norfolk & Norwich University Hospital
  • Arne Juette(Author)
    Norfolk & Norwich University Hospital
  • Reyer Zwiggelaar(Author)
Type Article
Original languageEnglish
Article number101880
Number of pages8
JournalArtificial Intelligence in Medicine
Volume107
Early online date29 May 2020
DOI
Publication statusPublished - 01 Jul 2020
View graph of relations
Citation formats

Abstract

In current breast ultrasound Computer Aided Diagnosis systems, the radi- ologist preselects a region of interest (ROI) as an input for computerized breast ultrasound image analysis. This task is time consuming and there is inconsistency among human experts. Researchers attempting to automate the process of obtaining the ROIs have been relying on image processing and conventional machine learning methods. We propose the use of a deep learn- ing method for breast ultrasound ROI detection and lesion localisation. We use the most accurate object detection deep learning framework – Faster- RCNN with Inception-ResNet-v2 – as our deep learning network. Due to the lack of datasets, we use transfer learning and propose a new 3-channel artificial RGB method to improve the overall performance. We evaluate and compare the performance of our proposed methods on two datasets (namely, Dataset A and Dataset B), i.e. within individual datasets and composite dataset. We report the lesion detection results with two types of analysis: 1) detected point (centre of the segmented region or the detected bounding box) and 2) Intersection over Union (IoU ). Our results demonstrate that the proposed methods achieved comparable results on detected point but with notable improvement on IoU. In addition, our proposed 3-channel artificial RGB method improves the recall of Dataset A. Finally, we outline some future directions for the research.

Documents

  • manuscript_R2

    Accepted author manuscript, 4.83 MB, PDF

    Embargo ends: 29 May 2021

    Request copy

    Licence: CC BY-NC-ND Show licence