Related Experiment Videos
Computerized detection of malignant tumors on digital mammograms
H Kobatake1, M Murakami, H Takeo
1Graduate School of Bio-Applications and Systems Engineering, Tokyo University of Agriculture & Technology, Koganei, Japan.
IEEE Transactions on Medical Imaging
|July 23, 1999
Summary
This study introduces an advanced tumor detection system for digital mammography. The system effectively identifies suspicious regions and characterizes malignant tumors, improving early cancer detection rates.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Oncology
Background:
- Digital mammography is crucial for breast cancer screening.
- Detecting subtle tumors with low contrast is a significant challenge.
- Accurate feature extraction is vital for differentiating malignant tumors.
Purpose of the Study:
- To develop an effective tumor detection system for digital mammography.
- To address the challenge of detecting low-contrast suspicious regions.
- To extract features that characterize malignant tumors.
Main Methods:
- A novel adaptive filter, the iris filter, was developed to enhance rounded opacities.
- Parameters reflecting boundary characteristics were proposed for tumor differentiation.
- A large-scale experiment with 1212 CR images was conducted to validate the system.
Main Results:
- The proposed system demonstrated high effectiveness in enhancing low-contrast tumors.
- Sensitivity of the tumor detection system reached 90.5%.
- The average number of false positives per image was low, at 1.3.
Conclusions:
- The developed tumor detection system is effective for digital mammography.
- The iris filter and boundary parameters contribute to improved detection accuracy.
- The system shows significant potential for enhancing breast cancer screening outcomes.