Multi-focus image fusion based on optimal defocus estimation. (August 2017)

• Record Type: Journal Article
• Authors: Aslantas, Veysel; Toprak, Ahmet Nusret
• Is Part Of: Computers & electrical engineering. Volume 62(2017)
• Page Start: 302
• Page End: 318
• Publication Date: 2017-08
• Subjects: Multi-focus image fusion -- Extended depth of field -- Defocus estimation; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2017.02.003
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: An optimal defocus estimation method is proposed to estimate the blur of multi focus images. An improved and efficient algorithm is presented to fuse the multiple focused images. The method making use of estimated defocus blur of images to detect focused pixels. The method outperforms state-of-the-art methods in both quantitative and visual comparisons. Abstract: One of the main drawbacks of the imaging systems is limited depth of field which prevents them from obtaining an all-in-focus image of the environment. This paper presents an efficient, pixel-based multi-focus image fusion method which generates an all-in-focus image by combining the images that are acquired from the same point of view with different focus settings. The proposed method first estimates the point spread function of each source image by utilizing the Levenberg–Marquardt algorithm. Then, artificially blurred versions of the source images are computed by convolving them with the estimated point spread functions. Fusion map is computed by making use of both the source and the artificially blurred images. At last, the fusion map is improved by morphological operators. Experimental results show that the proposed method is computationally competitive with the state-of-the-art methods and outperforms them in terms of both visual and quantitative metric evaluations. Graphical abstract: