BACKGROUND: Breast cancer is a tumor that begins in the breast tissue and is largely identified through X-ray imaging; however, human tissue, illumination, noise and other factors make the image's calcifications and masses unclear, which in turn affects the doctors' identification of lesions and normal tissue through X-ray imaging. Therefore, the rate of misdiagnoses can be reduced through the enhancement of X-ray images that make the images' calcifications and masses more prominent.
OBJECTIVE: Enhancing the breast image would highlight the calcifications and masses.
METHODS: One such way to do so is to use a curvelet that can detect curves and can, therefore, enhance the tumor characteristics. Essentially, existing methods perform a curvelet transform on each sub-image simultaneously; as the curvelet is based on the Radon transform, it involves complex computation and can easily result in difficulties. Based on this information, this article improved the algorithm that detects edges by curvelet and refines edges by wavelet. Simulation experiments using mammography X-ray images are implemented through Matlab.
RESULTS: The results suggest that, after implementation of the improved algorithm, the image's edges and textures are clear, the calcifications are independent, and there is no caking.
CONCLUSIONS: The curvelet method is imporved in efficacy with respect to the wavelet method.
Research articleFirst published online March, 2015
