Visualization of mammograms via fusion of enhanced features
Iztok Koren, Andrew F. Laine, Suzanne Smith, Edward L. Nickoloff, Fred J. Taylor · 1999
Image enhancement in mammography is typically concerned with either general visibility of all features or conspicuity of a specific sign of malignancy. We describe a synthesis of the two approaches through fusion of locally enhanced microcalcifications, circumscribed masses, and stellate lesions. Both local processing and image fusion are performed within a single wavelet transform framework which contributes to the computational efficiency of the method. The algorithm not only allows for efficient combination of specific features of importance, but also provides a flexible framework for incorporation of distinct enhancement methods and their independent optimization.