Abstract: A method can include receiving a low-resolution (LR) image, extracting a first feature embedding from the LR image, performing a first upsampling to the LR image by a first upsampling factor to generate a upsampled image, receiving a LR coordinate of a pixel within the LR image and a first cell size of the LR coordinate, generating a first residual image based on the first feature embedding, the LR coordinate, and the first cell size of the LR coordinate using a local implicit image function, and generating a first high-resolution (HR) image by combining the first residual image and the upsampled image via element-wise addition.

Abstract: An arbitrary-scale blind super resolution model has two designs. First, learn dual degradation representations where the implicit and explicit representations of degradation are sequentially extracted from the input low resolution image. Second, process both upsampling and downsampling at the same time, where the implicit and explicit degradation representations are utilized respectively, in order to enable cycle-consistency and train the arbitrary-scale blind super resolution model.

Abstract: A method for generating a high resolution image from a low resolution image includes retrieving a plurality of low resolution image patches from the low resolution image, performing discrete wavelet transform on each low resolution image patch to generate a first image patch with a high frequency on a horizontal axis and a high frequency on a vertical axis, a second image patch with a high frequency on the horizontal axis and a low frequency on the vertical axis, and a third image patch with a low frequency on the horizontal axis and a high frequency on the vertical axis, inputting the three image patches to a dual branch degradation extractor to generate a blur representation and a noise representation, and performing contrastive learning on the blur representation and the noise representation by reducing a blur loss and a noise loss.

Abstract: A system stores parameters of a feature extraction network and a refinement network. The system receives an input including a degraded image concatenated with a degradation estimation of the degraded image; performs operations of the feature extraction network to apply pre-trained weights to the input to generate feature maps; and performs operations of the refinement network including a sequence of dynamic blocks. One or more of the dynamic blocks dynamically generates per-grid kernels to be applied to corresponding grids of an intermediate image output from a prior dynamic block in the sequence. Each per-grid kernel is generated based on the intermediate image and the feature maps.

Abstract: According to a network space search method, an expanded search space is partitioned into multiple network spaces. Each network space includes a plurality of network architectures and is characterized by a first range of network depths and a second range of network widths. The performance of the network spaces is evaluated by sampling respective network architectures with respect to a multi-objective loss function. The evaluated performance is indicated as a probability associated with each network space. The method then identifies a subset of the network spaces that has the highest probabilities, and selects a target network space from the subset based on model complexity.