科研之友
微信
新浪微博
Facebook
分享链接
摘要
Medical image fusion processing is easy to produce artifact such as blurry edge or discontinuous edge will affect the performance of the fused image and prevent the doctors from achieve more accurate anatomic information. In this paper, a novel edge enhancement framework for image fusion method is proposed based on Nonsampled Shearlet transform (NSST) and Mathematical Morphology. The source images are decomposed into low-frequency and high-frequency bands; meanwhile, the edge saliency map is obtained by Mathematical Morphology edge detection. Varied fusion rules are applied for low-frequency and high-frequency bands. Integrating the edge saliency and the fused image based on inverse NSST, the final fused image captured more edge information. To verify the effectiveness of the approach, the experiments will be performed using six sets of source images and compared the results with other algorithms. Subjective and objective evaluation demonstrates that the performance of the proposed fusion algorithm has modified compared with the existing typical methods.
- 出版日期2018-6
- 单位河北大学; 中国人民解放军信息工程大学
