|
Scientific Visualization
| Issue Year: | 2016 |
| Quarter: | 4 |
| Volume: | 8 |
| Number: | 5 |
| Pages: | 83 - 93 |
|
| Article Name: |
FOCUSED REGION DETECTION USING MULTISCALE TOP-HAT TRANSFORM FOR MULTIFOCUS IMAGE FUSION |
| Authors: |
Defa Hu (China), Hailiang Shi (China), Weijin Jiang (China) |
| Address: |
Defa Hu
hdf666@163.com
Mobile E-business Collaborative Innovation Center of Hunan Province, Hunan University of Commerce, Changsha 410205, Hunan, China
Key Laboratory of Hunan Province for Mobile Business Intelligence, Changsha 410205, Hunan, China
Hailiang Shi
hlshi@zzuli.edu.cn
College of Mathematics & Information Science, Zhengzhou University of Light Industry, Zhengzhou 450002, Henan, China
Weijin Jiang
2858512981@qq.com
Mobile E-business Collaborative Innovation Center of Hunan Province, Hunan University of Commerce, Changsha 410205, Hunan, China
Key Laboratory of Hunan Province for Mobile Business Intelligence, Changsha 410205, Hunan, China |
| Abstract: |
A novel focused region detection algorithm based on multiscale top-hat transform is proposed, and based on that a multifocus image fusion method is put forward for visualization purpose. In the method, firstly the combined image features of the bright features and the dark features extracted by multiscale top-hat transform are compared to generate two initial decision maps as the focused regions, and then refine the maps by using morphological opening and closing iteratively until some conditions are met. Finally, the fused image is merged by copying the pixels of the focused regions with consistency verification on the boundary pixels to reduce the block effect. The proposed fusion method is tested on three datasets of multifocus images and compared with some traditional fusion methods, and experimental results demonstrate that the proposed method is effective to preserve the salient information and enhance the contrast for visualization, and superior to the traditional methods in terms of several quantitative evaluation measures. Moreover it is robust to mis-registration. |
| Language: |
English |
|
|
|
