Automated Detection of Masses in Digital Mammograms
Author Information
Author(s): Azlindawaty Mohd. Khuzi, Besar R, Wan Zaki WMD, Ahmad NN
Primary Institution: Multimedia University
Hypothesis
Can an automated system effectively identify masses in digital mammograms using texture analysis?
Conclusion
The proposed method using gray level co-occurrence matrices achieved a good detection rate for identifying masses in digital mammograms.
Supporting Evidence
- The method achieved a true positive rate of 96% for detecting masses.
- Otsu's method with GLCM features had an area under the ROC curve of 0.84.
- Using only four features, the system maintained a good detection rate.
- The study demonstrated that texture features can effectively differentiate between masses and non-masses.
Takeaway
This study created a computer program that helps doctors find lumps in breast images more easily and quickly.
Methodology
The study used digital mammogram images from the MIAS database, applying image processing techniques for enhancement, segmentation, and feature extraction using gray level co-occurrence matrices.
Potential Biases
Potential bias in the selection of images from the MIAS database.
Limitations
The study may not account for all variations in breast tissue density and mass characteristics.
Participant Demographics
The study used mammogram images without specific demographic details provided.
Statistical Information
P-Value
0.84
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website