Jove
Visualize
Contact Us

Related Experiment Video

Updated: May 8, 2026

Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
13:44

Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns

Published on: August 30, 2013

Automatic Tissue Classification for High-resolution Breast CT Images Based on Bilateral Filtering.

Xiaofeng Yang1, Ioannis Sechopoulos, Baowei Fei

  • 1Department of Radiology, Emory University.

Proceedings of Spie--The International Society for Optical Engineering
|September 13, 2013
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

High-speed optical tracking and augmented reality platform for image-guided interventions.

Journal of medical imaging (Bellingham, Wash.)·2026
Same author

Enhancing 1p/19q Classification in Brain Gliomas Using IDH Status: A Deep Learning Study.

AJNR. American journal of neuroradiology·2026
Same author

Bilateral Dynamic CT for Scapholunate Instability: Toward Personalized, Protocol-Standardized Kinematic Diagnosis.

Plastic and reconstructive surgery·2026
Same author

Polarized hyperspectral and polarized light microscopic imaging for enhanced visualization of white blood cells.

Journal of biomedical optics·2026
Same author

SAMIRA study on criteria for acceptability of medical radiological equipment (CARE). Survey results on European Union national legislations/regulations.

Physica medica : PM : an international journal devoted to the applications of physics to medicine and biology : official journal of the Italian Association of Biomedical Physics (AIFB)·2026
Same author

Medical hyperspectral imaging: an updated review of technology advancements and biomedical applications.

Journal of biomedical optics·2026
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

This article describes a new automated computer program designed to identify different types of breast tissue—skin, fat, and glands—within high-resolution 3D breast CT scans. By using specialized image-smoothing filters and fuzzy logic algorithms, the system accurately separates these tissues, even when images are noisy, helping doctors better assess breast density and patient risk.

Area of Science:

  • Medical imaging informatics within diagnostic radiology
  • Automated breast tissue classification systems in oncology

Background:

Accurate identification of breast components remains a significant challenge in diagnostic imaging. Prior research has shown that distinguishing glandular structures from surrounding skin is difficult due to overlapping intensity values. No prior work had resolved the need for fully automated segmentation in high-resolution dedicated breast computed tomography. That uncertainty drove the development of specialized filtering techniques to preserve anatomical boundaries. It was already known that noise reduction often blurs critical edges in medical scans. This gap motivated the exploration of advanced signal processing to maintain clarity. Previous studies relied heavily on manual intervention, which limits clinical throughput. The current landscape requires robust computational tools to improve diagnostic consistency across diverse patient populations.

Purpose Of The Study:

This study aims to develop an automated classification method for high-resolution dedicated breast computed tomography images. The primary objective involves accurately separating skin, fat, and glandular tissues to support clinical diagnostics. This gap motivated the need for quantitative measurements of breast composition and density distribution. The researchers seek to overcome limitations associated with manual segmentation, which is time-consuming and prone to variability. By implementing advanced filtering and classification algorithms, they intend to improve the efficiency of identifying high-risk patients. That uncertainty drove the team to create a system capable of handling noisy image data without losing anatomical detail. No prior work had resolved the specific challenge of distinguishing skin from glandular tissue when their intensity values overlap. The project focuses on creating a robust, reproducible workflow for clinical imaging environments.

Keywords:
Breast CTbias correctionbreast cancerbreast tissue classificationfuzzy C-Mean classificationimage classificationmultiscale filterimage segmentationfuzzy logicmedical imagingdiagnostic radiology

Frequently Asked Questions

The researchers utilize a dual-stage fuzzy C-mean algorithm. This approach first isolates the skin layer using morphological masks, then performs a second classification pass to distinguish between fatty and glandular components within the remaining volume.

The authors employ a multiscale bilateral filter. This tool is necessary to suppress image noise while simultaneously preserving sharp anatomical edges, which is critical for accurate segmentation of small structures like the skin layer.

Morphological operations are required to generate a skin mask. This step is essential because skin and glandular tissues exhibit similar intensity values in computed tomography scans, making intensity-based separation alone insufficient for accurate identification.

The researchers use Dice overlap ratios to quantify performance. This metric compares the automated segmentation output against manual labels provided by experts to determine the spatial agreement between the two datasets.

More Related Videos

Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
15:48

Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging

Published on: December 15, 2014

Related Experiment Videos

Last Updated: May 8, 2026

Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
13:44

Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns

Published on: August 30, 2013

Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
15:48

Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging

Published on: December 15, 2014

Main Methods:

Review approach involves a multi-step computational pipeline designed for high-resolution volumetric datasets. The investigators first apply a multiscale bilateral filter to stabilize signal intensity while protecting structural boundaries. They subsequently generate a skin mask through specific morphological operations based on spatial positioning. This step resolves intensity overlaps between superficial layers and internal structures. The team then executes a modified fuzzy C-mean algorithm in two distinct phases. Each phase targets specific tissue categories to optimize segmentation precision. They validate the performance by comparing automated outputs against manual expert delineations. Finally, the researchers assess algorithm stability by introducing synthetic noise into the original image sets.

Main Results:

Key findings from the literature indicate that the automated system achieves high spatial agreement with manual segmentation. The Dice overlap ratios for glandular tissue consistently exceeded 94.7% across the five patient datasets. The authors report that the dual-stage fuzzy logic approach effectively separates skin from fatty and glandular regions. This performance remains stable even when researchers introduce additional noise into the image files. The results demonstrate that the multiscale filtering successfully retains edge definition during the preprocessing stage. Quantitative measurements derived from this method provide reliable data on breast composition and density distribution. The evaluation confirms that the system is both accurate and resilient against common imaging artifacts. These findings establish a strong basis for automated tissue identification in dedicated breast scans.

Conclusions:

The authors propose that their automated segmentation framework provides a reliable alternative to manual image processing. Synthesis and implications suggest that the multiscale bilateral filtering approach effectively preserves anatomical boundaries while suppressing artifacts. The researchers demonstrate that their dual-stage fuzzy logic strategy successfully separates skin from internal glandular components. This study indicates that the proposed method maintains high performance even under conditions of increased image noise. The reported Dice overlap ratios confirm the accuracy of the automated classification relative to expert-led segmentation. These findings imply that the system could assist in quantifying breast density for clinical risk assessment. The authors conclude that their approach is both robust and suitable for high-resolution breast computed tomography datasets. Future clinical integration may benefit from the high precision observed across the five patient cases analyzed.

The system achieved Dice overlap ratios exceeding 94.7% for glandular tissue. This measurement was obtained by testing the algorithm on high-resolution scans collected from five distinct patients.

The authors propose that their method is robust and accurate for clinical use. They suggest that this automated system could provide quantitative measurements of breast composition, which is vital for identifying high-risk patients.