Detailed and enhanced multi-exposure image fusion using recursive filter
- PDF / 9,452,484 Bytes
- 22 Pages / 439.642 x 666.49 pts Page_size
- 28 Downloads / 215 Views
Detailed and enhanced multi-exposure image fusion using recursive filter Naila Hayat1 · Muhammad Imran2 Received: 24 May 2019 / Revised: 23 April 2020 / Accepted: 5 June 2020 / © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract A single photo is usually inadequate to represent a high-quality scene due to the dynamic range limitation. A high-quality image can be obtained by fusing multi-exposure images of the same scene. However, ghosting artifact can be produced in the fused image due to moving objects. To overcome this problem, we propose a detailed and enhanced multiexposure image fusion technique using an edge-preserving recursive filter. The proposed technique reduces the artifacts near edges and produces an HDR-like image without any ghosting artifact. The idea behind the proposed method is to first decompose the LDR multiple-exposed input images into the detail layer and the base layer to extract the sharp and fine details, respectively. To do so, first, the recursive filter is applied to input images. Then, these recursive-based output images are used for extracting the detail and base layer. Finally, the detail layer and the base layer are combined together to produce a detailed and enhanced image without artifacts. Additionally, the proposed method is suitable for multifocus image fusion. Experimental results prove the effectiveness of the proposed method over the existing methods both qualitatively and quantitatively. Keywords Multi-exposure · Low dynamic range · High dynamic range · Histogram · Recursive filter
1 Introduction Generally, the dynamic range of the real world scene is very large and the human eye can perceive a similar dynamic range. But the image sensors of digital cameras such as the charge coupled device (CCD) or complementary metal oxide semiconductor (CMOS) have limited dynamic range. Due to which, conventional digital cameras cannot represent the Muhammad Imran
[email protected] Naila Hayat nailahayat [email protected] 1
Sardar Bahdur Khan Women University, Quetta, Pakistan
2
Engineering & Management Sciences, Balochistan Univeristy of IT, Quetta, Pakistan
Multimedia Tools and Applications
image as natural and as comparable as the scene perceived by human visual system. As a result, the captured low dynamic range (LRD) scene may contain some dark areas; that is, underexposed regions and some bright areas; that is, overexposed regions [31, 38]. One way to get a high dynamic range image is to use multiple imaging devices and special camera sensors [1, 2, 32]. However, these solutions are very expensive and are not popular among common users. HDR imaging [15, 33] and multi-exposure image fusion (MEF) [10, 17, 20– 22, 25, 34, 35] are two other approaches to capture the high dynamic range images. These approaches are cheaper, hardware independent and easier to implement. The HDR image generated by the HDR imaging method cannot display directly on the LDR conventional devices. Therefore, tone mapping [6, 16] is used to map the HDR image into a low dyna
Data Loading...
