Denoising and enhancing digital mammographic images for visual screening

Area of Science:

• Medical Imaging
• Image Processing
• Radiology

Background:

• Dense regions in mammograms present challenges due to noise and low contrast.
• Effective screening of these regions is crucial for early detection of breast cancer.

Purpose of the Study:

• To develop an adaptive method for noise suppression and enhancement of dense regions in digital mammographic images.
• To improve the detection of subtle details like microcalcifications.

Main Methods:

• Image preprocessing to enhance local contrast and detail discrimination.
• Wavelet transform-based noise suppression and edge enhancement using Gaussian and Generalized Laplacian models.
• Adaptive shrinkage functions combined across scales, with a gain factor for detail enhancement.

Main Results:

• The method effectively suppresses noise and enhances edges in mammographic images.
• Improved local contrast and detection of microcalcifications and suspicious structures were observed.
• The process is adaptive, requiring fewer user parameter adjustments than other methods.

Conclusions:

• The proposed wavelet-based method offers significant improvements for mammographic image analysis, particularly in dense regions.
• It enhances the visibility of critical diagnostic features while minimizing artifacts.
• This technique shows promise for more accurate and efficient breast cancer screening.