Wavelet Decomposition in Laplacian Pyramid for Image Fusion

Abstract

The aim of image fusion is to combine information from the set of images to get a single image which contains a more accurate description than any individual source image. While the scene contains objects in different focus due to the limited depth-of-focus of optical lenses in camera then by using image fusion technique we can get an image which has better focus across all area. In this paper, a multifocus image fusion method using combination Laplacian pyramid and wavelet decomposition is proposed. The fusion process contains the following steps: first, the multifocus images are decomposed using Laplacian pyramid into several levels of pyramid. Then at each level of pyramid, wavelet decomposition is applied. The images at every level of wavelet are fused using maximum absolute value rule. The inverse wavelet transform is then applied to the combined coefficients to produce the fused image in laplacian pyramid. The final step is to reconstruct the combined image at every level of pyramid to get the fused image which shows an image retaining the focus from the several input images. Experimental results that are quantitatively evaluated by calculation of root mean square error, peak signal to noise ratio, entropy, and average gradient measures for fused image show the proposed method can give good result.

Fusion of Multi-Focus Images with Neighbour Local Distance

“…For comparison purposes, we performed fusion using methods: PCA method [1], Discrete Wavelet Transform (DWT) method [6], Laplacian Pyramid LP_PCA [15], LP_DWT [17] and Bilateral gradient (BG) [2].…”

“…In multi-scale fusion methods, the fusion process is performed on the source images after decomposing them into multiple-scales. The discrete wavelet transform (DWT) [4]- [9], Laplacian pyramid image fusion [10]- [17], Discrete cosine transform with variance calculation (DCT+var) [18], saliency detection based method (SD) [19] are examples of image fusion techniques under transformdomain.…”

Dempster-Shafer and Multi-Focus Image Fusion using Local Distance

“…For each image where we use different values of size of neighbourhood, , we define: (14) where is the source image, and is size of neighbourhood of local variability. We set the standard deviation of = for belongs to , we calculate: (15) for belongs to , we calculate: (16) for belongs to , we calculate: (17) This method obtains the information whether or not a pixel belongs to the focus area, for this we use the plausibility of which is the sum of the masses of the evidence for and the uncertainty :…”

“…On the other hand, the fusion by the methods at several scales is carried out on the source images after having decomposed them into several scales. As examples of these methods we cite among others: discrete wavelet transform (DWT) [4] - [7], the fusion of Laplacian pyramidal images [8] - [14], the discrete cosine transform with calculation of the variance (DCT + var) [15], the method based on the detection of salience (SD) [16].…”

Performance Evaluation of Quarter Shift Dual Tree Complex Wavelet Transform Based Multifocus Image Fusion Using Fusion rules

“…In literature various multi-resolution techniques are developed like the laplacian pyramid [4], gradient pyramid [5], discrete wavelet transform (DWT) [6], stationary wavelet transform (SWT) [7][8][9], multi-resolution singular value decomposition (MSVD) [10], discrete cosine harmonic wavelet transform (DCHWT) [11], lifting wavelet transform [12][13], double density discrete wavelet transform (DDDWT) [14] and Shearlet Transform [15]. The major problem with pyramid based methods is lack of spatial orientation selectivity, which results in blocking effect in the fused image.…”