| 摘要 |
Latest progresses in breast imaging using differential phase contrast technique pose the question how to fuse multiple information (yielded by the absorption, differential phase, and scattering signals) into a single, but more informative image for clinical diagnosis and evaluation. In this work, we propose an image fusion scheme based on the multiple-resolution (MR) framework. The three signals are first transformed into multiple bands presenting information at different frequency and then a two-step processing follows: (1) an intra-band processing enhances the local signal-to-noise ratio using a novel noise estimation method and context modeling; (2) an inter- band processing weights each band by considering their characteristics and contributions, as well as suppressing the global noise level. The fused image, which looks similar to conventional mammogram but with significantly enhanced detail features, is then reconstructed by inverse transform. This fused image is compatible with clinical settings and enables the radiologists to use their years of diagnosis experiences in mammography. |
