Improving the detection of simulated masses in mammograms through two different image-processing techniques.

Area of Science:

• Medical Imaging
• Radiology
• Image Processing

Background:

• Mammography is crucial for breast cancer detection.
• Dense breast tissue can obscure abnormalities, challenging mass detection.
• Image processing techniques aim to enhance lesion visibility.

Purpose of the Study:

• To evaluate if contrast-limited adaptive histogram equalization (CLAHE) or histogram-based intensity windowing (HIW) improves simulated mass detection in dense mammograms.
• To compare the efficacy of CLAHE and HIW against no processing and fixed windowing.

Main Methods:

• Simulated masses were embedded in digitized mammograms of dense breast tissue.
• Two image-processing methods, CLAHE and HIW, were applied with varied parameter settings.
• Image detection performance was assessed based on mass position, contrast, and processing parameters.

Main Results:

• The optimal histogram-based intensity windowing (HIW) setting significantly improved simulated mass detection compared to fixed windowing and no processing.
• Contrast-limited adaptive histogram equalization (CLAHE) did not show improved detection performance compared to no processing.
• Observer performance in the HIW experiment did not significantly differ between radiologists and non-radiologists.

Conclusions:

• Histogram-based intensity windowing (HIW) shows promise for improving mass detection in clinical mammograms and warrants further investigation.
• Contrast-limited adaptive histogram equalization (CLAHE) is unlikely to enhance mass detection in clinical mammograms.