Statistical performance analysis of a fast super-resolution technique using noisy translations

ABSTRACT:

The enrollment procedure is a key advance for super-goals (SR) remaking. An ever increasing number of gadgets allow to conquer this bottleneck utilizing a controlled situating framework, e.g., sensor moving utilizing a piezoelectric stage. This makes conceivable to gain various pictures of a similar scene at various controlled positions. At that point, a quick SR calculation can be utilized for effective SR reproduction. For this situation, the ideal utilization of r2 pictures for a goals upgrade factor r is for the most part insufficient to acquire fulfilling results because of the arbitrary error of the situating framework.

In this manner, we propose to take a few pictures around each reference position. We think about the blunder created by the SR calculation because of spatial vulnerability as a component of the quantity of pictures per position. We acquire a lower bound on the quantity of pictures that is important to guarantee a given blunder upper bound with likelihood higher than some coveted certainty level. Such outcomes give valuable indications to the outline of SR frameworks.

BASE PAPER: Statistical Performance Analysis

LEAVE A REPLY

Please enter your comment!
Please enter your name here