Detection and controlled manipulation of coherent structures in medical images

Much information in an image is concentrated in a few regions. These are dominated by structures characterized by discontinuities in color intensity. We present a wavelet based image analysis algorithm that identifies features that persist across the scales of the wavelet decomposition. This is achieved through the notion of sub-band combining which we introduce here. We first decompose the image into scales using the wavelet transform. Then, for each pixel, the combining process computes a function that collects all the information available at the sub-band (detail) scales. The result is a mask image indicating the significance of every point. Additional processing and image enhancement can now concentrate their efforts in regions where the masks show high values. Medical images, which are band-limited and contain no illumination artifacts, are especially suited to take advantage of sub-band combining. We show how our method can be employed to remove noise, enhance, and smooth features in images from MRI and mammography. Also, we allude to the use of this method to the creation of scale maps for level-of-detail computations