Gabor features have been widely used in texture classification and biometrics like face, fingerprint and palmprint recognition. Compared to the fruitful research of 2D Gabor wavelets, the research on 3D Gabor wavelets, is however, quite limited in literature. The project aims to explore the application of 3D Gabor wavelets for 3D image recognition and hyperspectral image classification problem, to find a feasible way to encode the relationships among different Gabor features in orientation, frequency and 3D spatial domains, to fuse the multi-view Gabor features with tensors and tries to apply the multi-task joint sparse representation for feature fusion. The research output of the project will benefit the Gabor wavelet theory and its application for 3D imaging techniques, and have significant impact on the progress of 3D biometrics,medical diagnosis and remote sensing