We create a dataset and a customized algorithm to run image super-resolution at 4K resolution in real time. Notably, the feature maps obtained from the Segment Anything are fused with the original images and fed into the network.
We used bilinear interpolation and JPEG compression (compression rate from random).
url: https://pan.baidu.com/s/1rCWIo5dzOfOUxdUXJ4vBIw?pwd=v907 password: v907
We propose the U-AsConvSR algorithm based on the AsConvSR: Fast and Lightweight Super-Resolution Network with Assembled Convolutions at CVPR 2023.