Implicit Transformer Network for Screen Content Image Continuous Super-Resolution
This addresses the need for high-resolution display of screen content in applications like remote cooperation and online education, but it is incremental as it adapts transformer-based techniques to a specific domain.
The paper tackles the problem of super-resolution for screen content images (SCIs), which existing methods designed for natural images fail to handle well, by proposing an Implicit Transformer Super-Resolution Network (ITSRN) that achieves high-quality continuous SR at arbitrary scales, significantly outperforming competitive methods on constructed SCI datasets.
Nowadays, there is an explosive growth of screen contents due to the wide application of screen sharing, remote cooperation, and online education. To match the limited terminal bandwidth, high-resolution (HR) screen contents may be downsampled and compressed. At the receiver side, the super-resolution (SR) of low-resolution (LR) screen content images (SCIs) is highly demanded by the HR display or by the users to zoom in for detail observation. However, image SR methods mostly designed for natural images do not generalize well for SCIs due to the very different image characteristics as well as the requirement of SCI browsing at arbitrary scales. To this end, we propose a novel Implicit Transformer Super-Resolution Network (ITSRN) for SCISR. For high-quality continuous SR at arbitrary ratios, pixel values at query coordinates are inferred from image features at key coordinates by the proposed implicit transformer and an implicit position encoding scheme is proposed to aggregate similar neighboring pixel values to the query one. We construct benchmark SCI1K and SCI1K-compression datasets with LR and HR SCI pairs. Extensive experiments show that the proposed ITSRN significantly outperforms several competitive continuous and discrete SR methods for both compressed and uncompressed SCIs.