Photos: Getting rid of camera wobble

In the blurry original on the left, it’s hard to detect details such as the wood grain of the towel rack, or the towel’s nap, that are readily\r\napparent in the sharpened version. The box around the woman’s face shows the rectangle selected for the deblurring process. Too small a rectangle\r\nlowers the result’s quality, but too large takes more time.

The deblurring technique alters color somewhat but clarifies many elements of this carving.

The deblurring method resolves details in feathers and a darker patch in the dark mask around the bird’s eye. However, the processed image also shows artifacts, visual echoes that look somewhat like ripples near sharp edges.

These pairs of images show close-ups of elements in original photos on the left, and the path of camera shake on the right. The shake path, called the blur kernel, is then used as the starting point for a process called deconvolution that can reverse some of the blur effects.

The deblurring algorithm helps even with relatively complicated images, such as this fountain.

The deblurring process can work well with text, which has plenty of high-contrast edges.

The deblurring process is complicated when there’s motion blur within the image (as in the case of this boy’s arm) as well as from the camera.