Optimizing the ROI (Region of Interest) Quality of Breast Cancer Skin Contour Images Using a Combination of Contrast Enhancement Methods Based on LiDAR Data

Abstract

Breast cancer has the highest prevalence and mortality rates among cancers in Indonesia, largely due to delayed diagnosis. One of the major challenges is the poor quality of medical images used for early detection. This study aims to improve the quality of images in the important area of the skin around breast cancer patients using Light Detection and Ranging (LiDAR) data by combining methods to enhance contrast and reduce noise. A total of 80 image data (40 cancer anomalies and 40 normal without cancer) and utilized MATLAB software version 24.1.0.2537033 (R2024a) for image processing, starting from increasing CNR and increasing SNR to semi-automatic ROI masking. The results showed that there was a significant increase in CNR values (an average of 38%) and SNR (an average of 42%). These results are supported by a paired T-test, which shows a significant difference between pre- and post-processed images, both in CNR α < 0.0001 and SNR α < 0.0001 parameters. These findings support the claim that image quality improvement is not only visually evident but also statistically evident. This study proves that the method used is effective in improving image quality and shows that LiDAR data has great potential in medical imaging systems.