Biological Image Fusion Using A SWT Based Variable-Weights Selection Scheme

A SWT based variable-weights selection scheme is proposed to merge the gray structural image and its corresponding color functional image for the localization of the protein in the biological community. After decomposing the original image by the stationary wavelet transform (SWT), multi-scaled transparencies are assigned for each pixel according to its SWT values. Then the fused product is generated by the inverse wavelet transform (ISWT) with multi-scaled transparencies. 117 groups of images are assessed by means of quantitative tests. Quantified indexes related with the fused product and original images are respectively calculated for the functional and nonfunctional area. Experimental results for several traditional image fusion methods indicate advantages of the proposed method.