Single-image motion deblurring using a low-dimensional hand shake blur model

A new algorithm for camera shake removal from a single image is presented. The motion blur point spread function (PSF) is represented in terms of the camera's angular velocity, which is in turn approximated by a linear function of time. Edge spread functions are estimated from isolated edges within the image, and the information along different directions is combined in an optimization procedure to find the best matching PSF. The image is then restored using the Wiener filter, followed by a post-processing algorithm for ringing artifact reduction. Results on real images taken with a digital camera are demonstrated.