Compound eye images super-resolution reconstruction based on Q-shift dual-tree complex wavelet transform

An image super-resolution reconstruction method based on Q-shift dual-tree complex wavelet transformation was proposed in this paper. After low resolution image sequence registration, the proposed method introduced MAP framework to complex wavelet domain. It used a generalized Gaussian wavelet prior model to provide accurate descriptions of the statistical distribution of complex wavelet coefficients in image data to obtain more directions of details. And then this method offered a new adaptive regularization technique and used the half-quadratic regularization iteration algorithm to determine the best regularization parameter and update image reconstruction simultaneity. This method offers good reconstruction results and global convergence. Experiment results indicate that without adding complexity, the result of this method is better than those of Joint MAP method and wavelet domain method by PSNR and subjective visual judgment.