1. A novel intuitionistic-near fuzzy sets based image fusion approach: development on hybrid MPI+OpenMP parallel model.
- Author
-
Biswas, Biswajit, Ghosh, Swarup Kr, and Ghosh, Anupam
- Subjects
IMAGE fusion ,FEATURE extraction ,FUZZY sets ,HIGH resolution imaging ,INFRARED imaging - Abstract
Image fusion is used to extract relevant features from different image modalities like infrared and visual images and to combine them into a single image effectively. In this article, we have introduced a hybrid-parallel intuitionistic-near set based fusion scheme through near feature map approach. The proposed hybrid-parallel fusion scheme fully utilizes distributed memory parallelism and OpenMP for shared-memory parallelism. First, the fuzzy image representation based intuitionistic fuzzy theory is considered. Second, the principal features in the infrared and visual images are mapped using near-fuzzy set. Third, final fusion features are measured from decomposed multiple image blocks through domain decomposition strategy and image features are extracted via a defined probe function. After that, the near features have been computed from both the original images via intuitionistic entropy-based probe function, the ultimate fusion image is achieved through perceptual threshold limit on the membership grades in the fuzzy space. Finally, the resultant fused image is obtained through defuzzification. A hybrid MPI and OpenMP model is adopted to reduce inter-node communication and parallelized codes. The experimental result shows that the proposed method effectively combines the relevant information of both source images and provides a high resolution image. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF