METHOD FOR REAL-TIME IMPLEMENTATION OF SUPER RESOLUTION

摘要

A method for real-time multi-frame super resolution (SR) of video content is provided. The method includes receiving a bitstream including an encoded video, motion metadata for a plurality of blocks of a frame of video content, and parameters. The motion metadata is estimated from the original video before downsampling and encoding. The motion metadata is averaged over consecutive blocks. The method includes upscaling the motion metadata for the plurality of blocks. The method also includes upscaling the decoded video using the upscaled motion metadata. The method also includes deblurring and denoising the upscaled video.

主权项

1. A method of increasing resolution using a real-time super-resolution (SR) technique, the method comprising:
receiving a bitstream including an encoded video sequence and metadata, wherein the metadata includes motion metadata for a plurality of blocks of a frame of video content and processing parameters, wherein the motion metadata is estimated from an original video before downsampling and encoding and wherein the motion metadata is downsampled by averaging over consecutive blocks; upscaling the motion metadata for the plurality of blocks; upscaling the decoded video with the help of upscaled motion metadata; and deblurring and denoising the upscaled video.