LAUSR

In Brazil, breast cancer is the disease with the highest mortality rate in women, mainly due to late diagnosis. The digital image processing has been used in order to provide both an improvement in the quality of the mammographic images as well as the evaluation of these images by radiologist. The goal of this work is to apply and evaluate (by an experienced radiologist) an adaptive histogram equalization algorithm and the variation of its parameters in mammographic digital images of dense breasts. We applied the Contrast-limited Adaptive Histogram Equalization (CLAHE) with different parameters (window size) in 98 images for the purpose of visual analysis and comparison with the respective original image. The peak signal-to-noise ratio (PSNR) ratio, the variance and Structural Similarity Index (SSIM) were calculated, as well as a visual evaluation of an experienced and specialized radiologist. The quantitative results showed the proximity between the averages of PSNR values for the set of windows tested. According to the radiologist assessment, the window size 15x15 provided a better contrast between fibroglandular tissue and adjacent structures. This study contributed to the contrast enhancement in dense mammograms, that is, belonging to those patients who present a higher risk of developing breast cancer. With the application of a simple mathematically and fast computational processing technique, it was possible to obtain better resultant images compared to the original ones, able to aid radiologists in a better diagnostic accuracy and in the early diagnosis of breast cancer.