As an essential ingredient of modern deep learning, attention mechanism, especially self-attention, plays a vital role in the global correlation discovery. However, is hand-crafted attention irreplaceable when modeling the global context? Our …
Our study aims to explicitly predict overlap relations and resolve overlaps in a differentiable way for the panoptic output.
We formulate the attention mechanism into an expectation-maximization manner and iteratively estimate a much more compact set of bases upon which the attention maps are computed.
We propose a novel neural network architecture to reduce streak artifacts generated in sparse-view 2D Cone Beam Computed To-mography (CBCT) image reconstruction.
We propose a novel deep network architecture based on deep convolutional and recurrent neural networksfor single image deraining.
We propose the Super-Resolution CliqueNet (SRCliqueNet) to reconstruct the high resolution (HR) image with better textural details in the wavelet domain.
We present an algorithm to address this problem by exploiting reliable edges and removing outliers in the intermediate latent images, thereby estimating blur kernels robustly