Adobe shows off new 'undo photo blur' feature
October 13, 2011 by Bob Yirka 
(PhysOrg.com) -- Despite all the advances in digital photography, most people are still plagued by the problem of blurry photos, a problem compounded by the use of cameras embedded in cell phones due to their small size. Problems of blurring generally can be divided into two types. The first is problems with focusing, which can usually be avoided if the camera operator will simply wait for the automatic focusing feature of their camera to do its job. The second type is much more difficult to solve as it involves camera movement while the image is being shot. It’s this second problem that Adobe has been working on as part of its Photoshop imaging software package. And based on a video shot by someone identified only as peterelst who posted it on Youtube, a recent demo of a new feature, or "sneak" as Adobe calls it, seems to indicate that they have made significant progress.
The technology behind the new feature (that may or may not actually wind up in Photoshop according to company reps) involves an algorithm based on the idea of blind deconvolution, which is where an iterative process is used to facilitate a point spread function. The idea is to calculate the speed at which the camera was moving when the picture was snapped so as to undo its effects. Or in other words, it attempts to reconstruct what the camera lens would have seen but for the movement. This is in stark contrast to current de-blurring functions in Photoshop and other image editing software which analyze an image looking for lines that form edges and bolsters them to make them appear sharper. The results with the new technique, at least in the demo, appear to be quite dramatic.
With the new feature, the photo is first loaded onto the computer and into the Photoshop type app, then some predefined parameters are loaded that more clearly define what sort of image is to be looked at. The image is then analyzed and a grayscale thumbnail (blur kernel) is displayed which shows how the image was blurred. Next a restore feature is activated and the blurred image is replaced with the newly sharpened image.
Despite this bit of theatrics by Adobe and lots of gushing by mainstream media seemingly intent on describing the new technology as the end of blurry photos, things are not quite as rosy as all that. This is because if the new feature is indeed added to Photoshop, it will still be out of most people’s grasp due to the high price of the product. Very few are likely to shell out hundreds of dollars to just to clear up a few images taken haphazardly on their cell phones or even their cameras. What really needs to happen is for this technology to be implemented in cameras so operators will never know they blurred their image by jiggling their camera in the first place.
This video is not supported by your browser at this time.
Adobe MAX 2011
© 2011 PhysOrg.com
-
From lemons to lemonade: Reaction uses carbon dioxide to make carbon-based semiconductor,
32 comments
-
Thioridazine kills cancer stem cells in human while avoiding toxic side-effects of conventional cancer treatments,
3 comments
-
SpaceX private rocket blasts off for space station (Update),
42 comments
-
Climate scientists say they have solved riddle of rising sea,
31 comments
-
SpaceX capsule has 'new car' smell, astronauts say (Update),
4 comments
-
Need a rigid insulation material???
19 hours ago
-
magnets or EMF in car bumpers to protect from fender bender
May 26, 2012
-
length of wire in a coil of known dimensions?
May 25, 2012
-
India Engineering Powerhouse
May 25, 2012
-
electromagnet core dereference between hard and soft iron
May 25, 2012
-
Measuring water pressure in an open tank
May 24, 2012
- More from Physics Forums - General Engineering
More news stories
Browser wars flare in mobile space
The browser wars are heating up again, but this time the fight is for dominance of the mobile Internet.
11 hours ago |
5 / 5 (1) |
3
Probability of contamination from severe nuclear reactor accidents is higher than expected: study
Catastrophic nuclear accidents such as the core meltdowns in Chernobyl and Fukushima are more likely to happen than previously assumed. Based on the operating hours of all civil nuclear reactors and the number ...
Technology / Energy & Green Tech
May 22, 2012 |
3.6 / 5 (22) |
56
|
HyperSolar shows dirty water no barrier to power world
(Phys.org) -- The Santa Barbara, California, company, HyperSolar, is set to transparently share the ups and downs of its research experiences toward the company’s ultimate vision, successfully producing ...
Technology / Energy & Green Tech
May 24, 2012 |
4.8 / 5 (16) |
17
|
SpotterRF debuts Radar Backpack Kit (w/ Video)
(Phys.org) -- SpotterRF has announced a special radar backpack kit designed to enhance situational awareness for soldiers on the ground. The company says its special radar is designed for warfighters as part ...
Technology / Hi Tech & Innovation
May 26, 2012 |
5 / 5 (5) |
13
|
Tesla to launch electric sedan in US on June 22
Tesla Motors said Tuesday it would begin deliveries of "the world's first premium electric sedan" on June 22, slightly ahead of schedule.
Technology / Energy & Green Tech
May 22, 2012 |
4.5 / 5 (12) |
18
Land and sea species differ in climate change response: study
(Phys.org) -- Marine and terrestrial species will likely differ in their responses to climate warming, new research by Simon Fraser University and Australia’s University of Tasmania has found.
'Unzipped' carbon nanotubes could help energize fuel cells, batteries
Multi-walled carbon nanotubes riddled with defects and impurities on the outside could replace some of the expensive platinum catalysts used in fuel cells and metal-air batteries, according to scientists at ...
T cells 'hunt' parasites like animal predators seek prey, study shows
By pairing an intimate knowledge of immune-system function with a deep understanding of statistical physics, a cross-disciplinary team at the University of Pennsylvania has arrived at a surprising finding: T cells use a movement ...
Computer model used to pinpoint prime materials for efficient carbon capture
When power plants begin capturing their carbon emissions to reduce greenhouse gases – and to most in the electric power industry, it's a question of when, not if – it will be an expensive undertaking.
Change in developmental timing was crucial in the evolutionary shift from dinosaurs to birds: study
At first glance, it's hard to see how a common house sparrow and a Tyrannosaurus Rex might have anything in common. After all, one is a bird that weighs less than an ounce, and the other is a dinosaur that ...
Nvidia trumpets Tegra 3 phone design wins for 2012
(Phys.org) -- Nvidia’s competitive war paint has a name, Tegra 3. On the heels of Nvidia announcements about lowering costs of its Tegra 3 processors and Nvidia-enabled tablets running Android Ice Cream ...
Oct 13, 2011
Rank: 4.8 / 5 (9)
Oct 13, 2011
Rank: 4.2 / 5 (5)
Oct 13, 2011
Rank: 3.8 / 5 (4)
This could be done with a relatively simple software update on (some) cameras.
Oct 13, 2011
Rank: 4 / 5 (4)
Good point, but what's really impressive about this algorithm is that it can figure out the motion vector from an already blurred photo. There are countless such sitting around in paper albums and computers. This could rescue historically important photos from the Civil War. Photos from the JFK assassination. This could help solve crimes by sharpening photos of crooks taken by cheap security cameras.
If the camera is still, an object blurred by moving through the frame also would have a motion vector, which could be clarified. Even a shaky video from a running person might be rescued, with enough cpu. Probabalistic algorithm?
Oct 13, 2011
Rank: 4 / 5 (2)
Oct 13, 2011
Rank: 3.3 / 5 (4)
Only if the movement was by the camera and not the subject.
Oct 13, 2011
Rank: 4 / 5 (2)
This algorithm could easily be given a highlighted section of an image and as far as it's concerned it was camera blur, doesn't matter, so long as the blur has a vector which can be determined and it can do it's sexy business it's all good.
For forensics this is sufficient, the point in a situation like that is extracting useful information about an individual and not creating aesthetically pleasing art.
So no, I don't think this algorithm would 'magically' stop working because of the blur coming from the camera or not, the algorithm is going to do what it's designed to do regardless of the reason behind a blur. Results may vary obviously.
Oct 13, 2011
Rank: not rated yet
First, what he got right: This deblurring tool is only for motion blur. The algorithm analyzes the distortion path and is able to reverse the process to clean up the picture. Namels is correct that it is primarily for camera movement. It will only work for a moving object if the object is moving uniformly like a car. You would still need to clean up the background afterwards. if you are jumping and waving your hands at the same time, it will do no good.
Now, for all you guys wanting this on a camera...
-First, it takes a huge amount of processing power not available on cameras.
-Second, there are professional cameras that have very little blur of any kind. You can already pay for a camera that takes better pictures on the front end.
-Third, there are higher end consumer cameras
-Continued-
Oct 13, 2011
Rank: not rated yet
Oct 13, 2011
Rank: 5 / 5 (1)
Also, If a picture is out of focus, you cannot resurrect it to the degree that the picture on top shows. The tools in photoshop are limited in that regard, because out of focus pictures are missing some amount of information, that does cause some artifacts. (This can be compensated for if you know the exact focal length and distance sometimes.)
This is an awesome tool for post processing, but there are already better things in place for higher end cameras. I don't think the author of the article has more than a passing knowledge in camera tech or photoshop.
Oct 13, 2011
Rank: not rated yet
Oct 13, 2011
Rank: 5 / 5 (3)
Oct 13, 2011
Rank: 5 / 5 (1)
Course, if processors continue to get faster on a per-unit-power-used basis at the current rate, then in *does quick math* less than 20 years this kind of thing could be built right into a camera, and used in real time with no noticeable lag. That's certainly not possible yet, but it will be.
Oct 14, 2011
Rank: 5 / 5 (1)
Yes. In order to restore information in a picture the information has to be there (as ina uniform blurring vectorfield where you can calculate the motion blur vector)
But information that is lost (i.e. by simple blur due to bad focus) cannot be restored - no matter what you do.
Oct 14, 2011
Rank: 5 / 5 (1)
Oct 14, 2011
Rank: not rated yet
Oct 22, 2011
Rank: not rated yet