Related Experiment Videos
Detecting film-screen artifacts in mammography using a model-based approach.
R Highnam1, M Brady, R English
1Medical Vision Laboratory, Engineering Science, Oxford University, U.K. rph@oxiva.com
IEEE Transactions on Medical Imaging
|January 11, 2000
Summary
This study developed an algorithm to distinguish breast cancer microcalcifications from artifacts caused by dust and dirt in mammograms. The new method achieved a 96% artifact detection rate, reducing false positives.
Area of Science:
- Medical Imaging
- Radiology
- Computational Pathology
Background:
- Microcalcifications are early indicators of breast cancer detected via mammography.
- Mammograms can produce false positives due to artifacts from dust and dirt mimicking microcalcifications.
Purpose of the Study:
- To develop and validate an algorithm for detecting film-screen artifacts in mammograms.
- To reduce false positives in mammographic screening by differentiating artifacts from true microcalcifications.
Main Methods:
- Utilized a model of the mammographic imaging process, focusing on inherent blurring functions.
- Employed advanced image processing techniques on carefully selected image representations.
- Tested the algorithm extensively to evaluate its performance in artifact detection.
Main Results:
- The algorithm achieved an artifact detection rate of approximately 96%.
- The developed method successfully identified no true microcalcifications as artifacts.
- Demonstrated significant potential for reducing false-positive results in mammography.
Conclusions:
- The algorithm effectively detects dust and dirt artifacts in mammograms.
- This approach enhances the accuracy of mammographic screening by minimizing false positives.
- Improved artifact detection contributes to more reliable breast cancer diagnosis.