Related Experiment Video
Updated: Jun 9, 2026
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
Radial-searching contour extraction method based on a modified active contour model for mammographic masses
Toshiaki Nakagawa1, Takeshi Hara, Hiroshi Fujita
1Department of Intelligent Image Information, Division of Regeneration and Advanced Medical Sciences, Graduate School of Medicine, Gifu University, Gifu, Japan. nakagawa@fjt.info.gifu-u.ac.jp
Radiological Physics and Technology
|September 8, 2010
Summary
This study introduces an automated method for accurately outlining masses on digital mammograms, enhancing computer-aided diagnosis (CAD) systems. The novel radial-searching contour extraction improves mass segmentation accuracy for better diagnostic support.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Image Segmentation
Background:
- Accurate segmentation of masses in digital mammograms is crucial for improving computer-aided diagnosis (CAD) systems.
- Existing methods may struggle with precise contour recognition, particularly for masses with complex shapes.
Purpose of the Study:
- To develop an automated extraction scheme for precise recognition of mass contours on digital mammograms.
- To enhance the accuracy and utility of CAD systems through improved mass segmentation.
Main Methods:
- A modified active contour model (ACM) with a radial-searching contour extraction technique was proposed.
- The method determines the mass center via density gradient analysis and constrains contour movement radially.
- Image forces were calculated using an edge-intensity image and a degree-of-separation image for enhanced accuracy.
Main Results:
- The automated method was tested on 53 digitized mammograms with 53 masses exhibiting difficult contours.
- Quantitative comparison with physician-drawn segmentations showed high overlap ratios.
- 30 cases achieved >81% overlap, and 45 cases achieved >61% overlap with correct segmentation.
Conclusions:
- The developed automated mass contour extraction technique shows significant promise for improving CAD systems.
- The method offers a reliable approach for segmenting masses, especially those with challenging boundaries.
- This technique can contribute to more accurate and efficient analysis of mammographic images.
