Implicit Transformer Network for Screen Content Image Continuous Super-Resolution (Supplementary Materials) Jingyu Y ang

Figure 1 presents some examples of our dataset. All the four datasets do not contain personally identifiable information or offensive content. Figure 1 presents the visual comparison results for the compressed images. One reason is that we introduce a learnable weighting factor in implicit position encoding. SR performance for natural images.All the models(except TTSR) are On the B100 dataset, our method is only 0.06 dB less than the best method NLSN.