How AI describes TecoGAN

TecoGAN is a project that implements a temporally coherent GAN for video super-resolution. The repository provides code for training and inference, data preparation scripts, and pre-trained models to generate high-resolution video frames with temporal consistency. Developed by Mengyu Chu, You Xie, Laura Leal-Taixe, and Nils Thürey at Technical University of Munich, the work was published in ACM Transactions on Graphics.