3D segmentation and reconstruction of endobronchial ultrasound

Xiaonan Zang, Mikhail Breslav, William Evan Higgins

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Scopus citations


State-of-the-art practice for lung-cancer staging bronchoscopy often draws upon a combination of endobronchial ultrasound (EBUS) and multidetector computed-tomography (MDCT) imaging. While EBUS offers real-time in vivo imaging of suspicious lesions and lymph nodes, its low signal-to-noise ratio and tendency to exhibit missing region-of-interest (ROI) boundaries complicate diagnostic tasks. Furthermore, past efforts did not incorporate automated analysis of EBUS images and a subsequent fusion of the EBUS and MDCT data. To address these issues, we propose near real-time automated methods for three-dimensional (3D) EBUS segmentation and reconstruction that generate a 3D ROI model along with ROI measurements. Results derived from phantom data and lung-cancer patients show the promise of the methods. In addition, we present a preliminary image-guided intervention (IGI) system example, whereby EBUS imagery is registered to a patient's MDCT chest scan.

Original languageEnglish (US)
Title of host publicationMedical Imaging 2013
Subtitle of host publicationUltrasonic Imaging, Tomography, and Therapy
StatePublished - 2013
EventMedical Imaging 2013: Ultrasonic Imaging, Tomography, and Therapy - Lake Buena Vista, FL, United States
Duration: Feb 12 2013Feb 14 2013

Publication series

NameProgress in Biomedical Optics and Imaging - Proceedings of SPIE
ISSN (Print)1605-7422


OtherMedical Imaging 2013: Ultrasonic Imaging, Tomography, and Therapy
Country/TerritoryUnited States
CityLake Buena Vista, FL

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Atomic and Molecular Physics, and Optics
  • Radiology Nuclear Medicine and imaging
  • Biomaterials


Dive into the research topics of '3D segmentation and reconstruction of endobronchial ultrasound'. Together they form a unique fingerprint.

Cite this