deep learning

We propose a novel approach for single image super-resolution to simultaneously predict high resolution images and the degradation kernels.

We propose a high-fidelity differentiable renderer for point clouds. We demonstrate how the proposed technique can be used to leverage contemporary deep neural networks to achieve state-of-the-art results in challenging geometry processing tasks.

Representing and generating shapes using neural networks

We present a detail-driven deep neural network for point set upsampling. A high-resolution point set is essential for point-based …

Making use of this extremely flexible yet unstructured form of shape represenation.

We set a new benchmark for single-image super-resolution by exploiting progressiveness both in architecture and training. The proposed multi-scale models, ProSR and ProSRGan, improve the reconstruction quality in terms of PSNR and visual quality respectively. ProSR is one of the winning teams.

Pushing the limit of “zoom-in"s.