Microsoft researchers presented a novel new anti-blurring technology at the recent SIGGRAPH 2010 conference in Los Angeles. The solution is fundamentally simple—using a hardware attachment on a camera to measure the movement produced by an errant hand shake or other disruption.
This movement data, combined with an in-camera analysis of the picture itself, allows a device to come up with a stronger approximation of the corrections needed to "sharpen," or deblur, a picture back to its intended look.
The four researchers named in the study managed to construct their hardware sensor package completely off-the-shelf, using a combination of one three-axis accelerometers, three gyroscopes, and a Bluetooth radio all wired to an open-source Arduino controller.
"Our method is completely automatic, handles per-pixel, spatially-varying blur, and out-performs the current leading image-based methods," reads the accompanying paper.
"To the best of our knowledge, this is the first work that uses 6 [degrees-of-freedom] inertial sensors for dense, per-pixel spatially-varying image deblurring and the first work to gather dense ground-truth measurements for camera-shake blur."
The team strapped their large sensor attachment to the bottom of a Canon 1Ds Mark III digital SLR camera, though they note that a smaller attachment than the existing prototype could find a future as an aftermarket camera accessory.
That's not to say that the group's work is the death-knell for all camera-based image blur. Although highly improved versus any kind of manual digital sharpening process, the deblurred images can still suffer a slight loss of quality compared to a pristine, non-blurry original.
The research group has set up a page featuring mouseover-based image comparisons that highlight the results of their technology versus other blur-removing methods. As well, the full text of the paper created by Neel Joshi, Sing Bing Kang, C. Lawrence Zitnick, and Richard Szeliski can be found at the accompanying Microsoft Research site.
This movement data, combined with an in-camera analysis of the picture itself, allows a device to come up with a stronger approximation of the corrections needed to "sharpen," or deblur, a picture back to its intended look.
The four researchers named in the study managed to construct their hardware sensor package completely off-the-shelf, using a combination of one three-axis accelerometers, three gyroscopes, and a Bluetooth radio all wired to an open-source Arduino controller.
"Our method is completely automatic, handles per-pixel, spatially-varying blur, and out-performs the current leading image-based methods," reads the accompanying paper.
"To the best of our knowledge, this is the first work that uses 6 [degrees-of-freedom] inertial sensors for dense, per-pixel spatially-varying image deblurring and the first work to gather dense ground-truth measurements for camera-shake blur."
The team strapped their large sensor attachment to the bottom of a Canon 1Ds Mark III digital SLR camera, though they note that a smaller attachment than the existing prototype could find a future as an aftermarket camera accessory.
That's not to say that the group's work is the death-knell for all camera-based image blur. Although highly improved versus any kind of manual digital sharpening process, the deblurred images can still suffer a slight loss of quality compared to a pristine, non-blurry original.
The research group has set up a page featuring mouseover-based image comparisons that highlight the results of their technology versus other blur-removing methods. As well, the full text of the paper created by Neel Joshi, Sing Bing Kang, C. Lawrence Zitnick, and Richard Szeliski can be found at the accompanying Microsoft Research site.
source: http://www.pcmag.com/article2/0,2817,2367240,00.asp
 | ||
| Before |
 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| After |
No comments:
Post a Comment