Enhancing mammography: a comprehensive review of computer methods for improving image quality

Area of Science:

• Radiology and Medical Imaging
• Artificial Intelligence in Healthcare

Background:

• Mammography is crucial for breast cancer detection but faces image quality challenges.
• Poor image quality can result in misdiagnosis, increased radiation, and higher costs.

Purpose of the Study:

• To review traditional and machine learning methods for mammography image enhancement.
• To identify effective techniques for improving diagnostic accuracy and reducing errors.

Main Methods:

• Literature search (2015-2024) of 115 articles on contrast enhancement and noise reduction.
• Evaluation of traditional methods (histogram equalization, filtering) and machine learning (ML) approaches.
• Focus on ML architectures like denoising autoencoders with convolutional neural networks (CNNs).

Main Results:

• Machine learning, especially autoencoder-CNN hybrids, effectively enhances mammography image quality without losing detail.
• Reviewed techniques show success in improving visual quality of mammograms.
• Identified persistent challenges include high noise, inconsistent metrics, and limited datasets.

Conclusions:

• ML-based image enhancement offers significant potential for improving mammography.
• Further research is needed to overcome current limitations and advance the field.
• Addressing noise, standardizing metrics, and expanding datasets will enhance clinical utility.