Image Fusion Method Based on Short Support Symmetric Non-Separable Wavelet

Abstract

In this paper, image fusion method based on a new class of wavelet — non-separable wavelet with compactly supported, linear phase, orthogonal and dilation matrix [Formula: see text] is presented. We first construct a non-separable wavelet filter bank. Using these filters, the images involved are decomposed into wavelet pyramids. Then the following fusion algorithm was proposed: for low-frequency part, the average value is selected for new pixel value, For the three high-frequency parts of each level, the standard deviation of each image patch over 3×3 window in the high-frequency sub-images is computed as activity measurement. If the standard deviation of the area 3×3 window is bigger than the standard deviation of the corresponding 3×3 window in the other high-frequency sub-image. The center pixel values of the area window that the weighted area energy is bigger are selected. Otherwise the weighted value of the pixel is computed. Then a new fused image is reconstructed. The performance of the method is evaluated using the entropy, cross-entropy, fusion symmetry, root mean square error and peak-to-peak signal-to-noise ratio. The experiment results show that the non-separable wavelet fusion method proposed in this paper is very close to the performance of the Haar separable wavelet fusion method.

A modified statistical approach for image fusion using wavelet transform

“…3 is applied on a set of multi-focus and multi-spectral images. For the purpose of comparison and to show the effectiveness of the proposed method, the same set of input images are subjected to image fusion algorithm, used in [1], but using separable orthogonal wavelet transform, such as Daubechies tap 4 filter, called hereafter as old method where the LL or approximation sub bands of DWT decomposed images are fused by a simple pixel-by-pixel averaging technique. But in our proposed method, approximation sub bands are fused using information from corresponding detail sub bands following the modified algorithm discussed in Sect.…”

“…The input images are first subjected to two levels of DWT decomposition. The wavelet coefficient in low frequency sub-bands, i.e., either LL 1 2 (of the first image) or LL 2 2 (of the second image) is chosen based on the combined edge information in the corresponding high frequency sub-bands, i.e., LH 1 2 , HL 1 2 and HH 1 2 or LH 2 2 , HL 2 2 and HH 2 2 . Here, mean and standard deviation over 3 × 3 windows are used as activity measurement to find the edge information present in the high frequency sub-bands.…”

“…Table 1 shows the performance of the fusion of multi-focus images, where as the old method indicates the algorithm proposed in [1]. Theoretically, the Entropy and PSNR of the fused image should be high whereas the fusion symmetry should be small.…”

“…Compared with the input images, the new image contains more comprehensive, more accurate and more stable information. Therefore, with the availability of multi-sensor data in many fields such as remote sensing, medical imaging, machine vision and military applications, the multi-sensor image fusion has emerged as a new and promising research area [1].…”

Multi-Source Multi-Sensor Image Fusion Based on Bootstrap Approach and SEM Algorithm

“…Specialists, who analyze images, use multi-detector manners, to fuse several sets of data, in order to extract the best possible information of a target or a region [5]. In recent years, image fusion approaches and their applications have been widely investigated [6][7][8][9]. Some researchers focused their attention on studying probabilistic based image fused algorithm.…”