Affordable Access

deepdyve-link
Publisher Website

Vision-Based Surgical Field Defogging.

Authors
  • Luo, Xiongbiao1
  • McLeod, A Jonathan2
  • Pautler, Stephen E3
  • Schlachta, Christopher M3
  • Peters, Terry M2
  • 1 Department of Computer Science, Fujian Key Laboratory of Sensing and Computing for Smart City, Xiamen University, Xiamen, China. , (China)
  • 2 Robarts Research Institute, Western University, London, ON, Canada. , (Canada)
  • 3 Departments of Surgery and Oncology, Western University, London, ON, Canada. , (Canada)
Type
Published Article
Journal
IEEE Transactions on Medical Imaging
Publisher
Institute of Electrical and Electronics Engineers
Publication Date
Oct 01, 2017
Volume
36
Issue
10
Pages
2021–2030
Identifiers
DOI: 10.1109/TMI.2017.2701861
PMID: 28504934
Source
Medline
License
Unknown

Abstract

Fogged surgical field visualization that is a common and potentially harmful problem can lead to inappropriate device use and incorrectly targeted tissue and increase surgical risks in endoscopic surgery. This paper aims to remove fog or smoke on endoscopic video sequences to augment and maintain a direct and clear visualization of the operating field. A new visibility-driven fusion defogging framework is proposed for surgical endoscopic video processing. This framework first recovers the visibility and enhances the contrast of hazy images. To address the color infidelity problem introduced by the visibility recovery, the luminances of the recovered and enhanced images are fused in the gradient domain, and the fused luminance is reconstructed by solving the Poisson equation in the frequency domain. The proposed method is evaluated on clinical videos that were collected from prostate cancer surgery. The experimental results demonstrate that the proposed framework defogs endoscopic images more robustly than currently available methods. Additionally, our method also provides an effective way to improve the visual quality of medical or high-dynamic range images.

Report this publication

Statistics

Seen <100 times