Image Fusion Based on Multi-wavelet Transform

Multi-wavelet is an extension from wavelet theory, and it has several particular advantages over scalar wavelet on image processing since the multi-wavelet analysis can offer a more precise method for image analysis than wavelet multi-resolution analysis can do. In this paper, two new image fusion algorithms are presented: One is based on image multi-wavelet decomposition and fractal dimension, and the other is based on image multi-wavelet decomposition and gradient characteristic domain mask. Based on the multi-wavelet transform with different scales, different source images can be separately decomposed. Here, the corresponding-level high frequency images are merged with fractal dimension as weight or gradient characteristic domain mask as weight while the corresponding-level lowest frequency images are merged with local energy as weight. Experiments are carried out using multi-focus image, visual light image and infrared image separately as two source images with results showing the two algorithms are feasible and effective