Multiscale Morphological Enhancement of Mammographic Images Using Top-Hat Reconstruction

Accurate interpretation of mammographic images is essential for the early detection and diagnosis of breast cancer. However, the quality of these images is often affected by noise, low contrast, and resolution limitations, which can hinder the identification of relevant structures such as microcalcifications and may compromise clinical interpretation. In this work, a mammographic image enhancement method based on the Multiscale Top-Hat Transform by Reconstruction (MTHR) is proposed. This multiscale morphological approach enables the extraction and enhancement of relevant features through morphological reconstruction operations that identify bright and dark regions of the image, thereby improving contrast and the visibility of fine details while preserving structural information. The proposed method was evaluated using 24 mammographic images from a public dataset and compared with several state-of-the-art contrast enhancement algorithms, including MSTH, QHELC, CLAHE, HE, and MMBEBHE. Experimental results, assessed using PSNR, entropy, AMBE, and REC metrics, show that the proposed method achieves competitive improvements in image quality and visibility of relevant structures compared with existing methods.