An image fusion algorithm based on lifting wavelet transform

The directional characteristic of the low-frequency and high-frequency coefficients based on the wavelet transform for original images is discussed and analyzed, and a novel image fusion algorithm based on the lifting wavelet transform is proposed in this paper. Firstly, the source images are transformed to the frequency domain by means of the lifting wavelet. Then, the resultant coefficients of the low-frequency sub-band are achieved by comparing the covariance of the coefficients of different images. Meanwhile, the resultant coefficients of each high-frequency sub-band are calculated according to the matching measure between the directional characteristic of the coefficients in the same sub-band and the quad-tree structure relationship of the coefficients with the same direction in different sub-bands. At last, the fusion-resultant image is obtained through the reversely lifting wavelet transformation. Several evaluation indexes, such as entropy, average grads, PSNR and RMSE, are employed to judge the experimental images with different fusion methods. The comparison results show that the proposed method image fusion algorithm based on the lifting wavelet transform is better than conventional methods, and has much application value.