Joint Adaptive Regularization and Thresholding on Manifolds for Depth Restoration from RGB-D Data PROJECT TITLE : Depth Restoration From RGB-D Data via Joint Adaptive Regularization and Thresholding on Manifolds ABSTRACT: By integrating the properties of local and non-local manifolds that offer low-dimensional parameterizations of local and non-local geometry of depth maps, we have developed a novel depth restoration algorithm using RGB-D data. Manifold regularisation is presented to enhance smoothing along the manifold structure by first defining a local manifold model that favours local nearby relationships of pixels in depth. It is also possible to exploit the patch-based manifold's non-local properties, such as its self-similar structures, to develop highly data-adaptive orthogonal bases to extract extended visual patterns. A manifold thresholding operator in 3D adaptive orthogonal spectral bases (eigenvectors of the discrete Laplacian of local and non-local manifolds) is further defined to keep only low graph frequencies for depth map restoration. Lastly, we present an efficient alternating direction approach of multipliers optimization framework that combines adaptive manifold regularisation and thresholding to solve the inverse problem of depth map recovery... Our strategy outperforms the current state-of-the-art in both objective and subjective quality evaluations, according to the findings of experiments. Did you like this research project? To get this research project Guidelines, Training and Code... Click Here facebook twitter google+ linkedin stumble pinterest A Fusion Approach to Infrared and Visible Images with DenseFuse Joint Color-Guided Internal and External Regularizations for Depth Super-Resolution