Comparison of two computationally feasible image deblurring techniques for smartphones

This paper presents a comparison of two computationally efficient motion deblurring techniques that have been shown to be practically feasible for deployment on smartphones. One technique involves the use of smartphone inertial sensors to provide an estimate of blur causing point-spread-function and the other involves the use of a short-exposure image that is obtained electronically at the same time the blurred image is captured. Both objective and subjective evaluations are carried out for a comprehensive comparison of these two techniques. It is shown that on average the technique utilizing a short-exposure image leads to more effective deblurrring outcomes compared to the technique utilizing inertial sensors.