Super resolution using multi resolution inputs

In this result we demonstrate that our method is able to handle input sequences of varying space-time resolutions. We provide two high resolution images (320x240) with a time gap of 1.6 seconds from a still camera and a low resolution video of 1.6 seconds duration with a spatial resolution of 160x120, and generate a high resolution (spatial) video of 50 frames. The reconstructed output shows high spatial resolution.

Input Images

Input video (385 Kb AVI)

Interpolated frame from the input video

Output video (385 Kb AVI)

Corresponding super resolved frame with increase spatial resolution