New Photo-Deblurring Research at 33rd Siggraph Conference

Soon, you won’t have to delete your blurry pictures, thanks to a new process of deblurring images revealed by MIT and University of Toronto researchers. This is good news because as mobile phones and digital cameras become smaller and lighter, even the smallest hand jitters will distort your pictures.

The MIT-U of Toronto team unveiled their technique for “Removing Camera Shake from a Single Photograph” at the 33rd Annual Siggraph Conference in Boston. They did it in a seminar called “Removing Camera Shake from a Single Photograph” (don’t you love how scientific papers have such creative titles?).

Research papers are not exactly light reading, so if esoteric mathematics makes you break out in rashes, then you can just look at the pretty picture below from DigitalCameraInfo.com:

How is this possible? Well, a motion-blurred picture is actually (deep down inside) a clear picture, but the camera lens moved while the picture was being taken, so a blur got “overlapped” on the sharp picture (math geniuses would get all excited about this and call it a “convolution,” but we’ll spare you the details). To make a long story short: the MIT-U of Toronto researchers created a “deconvolution” technique. By calculating the blur, they can sort of subtract it to reveal a clear picture.

This is great because not everybody wants to carry around a huge tripod just to take pictures using a mobile phone. Unfortunately, it will take a while (maybe two years) before the technique is ready for software for the general market. Also, the process does not correct other kinds of blur (like when the lens is out of focus, when an object is moving too fast for the camera’s shutter speed, or when trying to capture motion in pictures). That said, this nifty new process is a most welcome solution to a bothersome problem.

Via DigitalCameraInfo.com

Soon, you won’t have to delete your blurry pictures, thanks to a new process of deblurring images revealed by MIT and University of Toronto researchers. This is good news because as mobile phones and digital cameras become smaller and lighter, even the smallest hand jitters will distort your pictures.

The MIT-U of Toronto team unveiled their technique for “Removing Camera Shake from a Single Photograph” at the 33rd Annual Siggraph Conference in Boston. They did it in a seminar called “Removing Camera Shake from a Single Photograph” (don’t you love how scientific papers have such creative titles?).

Research papers are not exactly light reading, so if esoteric mathematics makes you break out in rashes, then you can just look at the pretty picture below from DigitalCameraInfo.com:

How is this possible? Well, a motion-blurred picture is actually (deep down inside) a clear picture, but the camera lens moved while the picture was being taken, so a blur got “overlapped” on the sharp picture (math geniuses would get all excited about this and call it a “convolution,” but we’ll spare you the details). To make a long story short: the MIT-U of Toronto researchers created a “deconvolution” technique. By calculating the blur, they can sort of subtract it to reveal a clear picture.

This is great because not everybody wants to carry around a huge tripod just to take pictures using a mobile phone. Unfortunately, it will take a while (maybe two years) before the technique is ready for software for the general market. Also, the process does not correct other kinds of blur (like when the lens is out of focus, when an object is moving too fast for the camera’s shutter speed, or when trying to capture motion in pictures). That said, this nifty new process is a most welcome solution to a bothersome problem.

Via DigitalCameraInfo.com