Image fusion of ultrasound tomography based on nonsubsampled contourlet transform and image segmentation

In recent years there has been considerable interest in ultrasound tomography imaging technology for breast cancer diagnosis. Transmission image and reflective image can be reconstructed in ultrasound tomography. Clear tissue reflective interface can be provided in reflective images and distinguished inner information of the tissue can be provided in transmission images. A nonsubsampled contourlet transform and image segmentation based image fusion approach is presented for ultrasound tomography using transmission image and reflective image. The experiments demonstrate that a fused result image with clear tissue reflective interface and enough tissue inner information can be obtained employing the proposed image fusion approach. The experiments using some image fusion metrics to assess the performance of fusion approaches show that the proposed fusion approach can provide best fusion performance for ultrasound tomography than other commonly used fusion approaches.