'Ghost images' may be the future of seeing our skeletons.
Looking inside ourselves is never easy, especially when we use radiation. A new x-ray method could be cheaper and less harmful. (Photo: toeytoey/Shutterstock)
While X-rays have helped us diagnose many medical ailments, there are lingering concerns about the amount of radiation required for the procedure.
The cancer risks related to repeated exposure for medical imaging are considered slight, but experts don't really know for sure. In the vein of better safe than sorry, we've been hunting for improvements in medical imaging, and a group of researchers from the Chinese Academy of Sciences in Beijing think they have a set-up that could eventually take clear X-ray images using less radiation.
"Our system is much smaller and cheaper, and it could even be portable if you needed to take it into the field," Wu Ling-An, a physicist involved in the project, told Science magazine.
A spooky way to see
The project, detailed in the journal Optica, modifies a technique called ghost imaging.
Ghost imaging produces an image of an illuminated object in what seems like a roundabout way. Imagine two transparent boxes hooked up to a computer. One of these boxes is empty and its back wall has a grid of many pixels on it, like a camera. The second box contains an object and it only has one pixel along its back wall. (If you tried to take a picture of something using only a camera with only one pixel, you'd end up with a black, white or gray square.)
Now shine a laser into a beam splitter so one beam enters the empty box and the other enters the box with the object in it. Make sure that each of the beams hit the same spot in their respective boxes. So the beam in the empty box hits the top left pixel while the beam in the box with the object hits the top left corner.
Now, move the laser beam so each pixel in the empty box is hit at the same time as the corresponding spot in the box with the object in it. Obviously, the beam entering the box with the object in it will hit that object at certain points and not reach the back wall of the box. The processor that the boxes are hooked up to will only register signals when both beams hit their respective back walls. The result is that the processor ends up with a silhouette of the image in the second box even though the box with all the pixels along the back wall has no access to the object in the box with a single pixel on its back wall.
If it's still a little unclear, this TEDx talk below may help.
Wu and the rest of the researchers aren't the first to use ghost imaging to make X-rays. However, the first ghost image X-rays needed a "building-size source of intense x-rays"; the Chinese team managed to do it on top of a table.
Their system fired X-rays first through a copper plate to block out any unwanted beams. The X-rays traveled through that copper plate and struck a rotating piece of sandpaper. Behind that rotating piece of sandpaper was a camera taking photographs of the sandpaper as the X-rays went through. A computer recorded where the X-rays came through in each part of the sandpaper, capturing more than 10,000 frames.
After this, the researchers reset the sandpaper to its original starting position, placed an object — in this case, a sea shell and a stainless steel mask with the letters CAS stenciled onto it — exactly where the first camera was located. They then placed another X-ray camera 10 centimeters behind the object, though the researchers noted that the distances didn't matter so long as the entire beam could be caught. X-rays were then fired through the copper plate, through the sandpaper and onto the sea shell and steel mask.
The result was ghost image X-rays for both objects, including a smudgy but legible view of the stenciled letters.
To be considered viable, their method would have to demonstrate that the total X-ray dose — which, in this case, requires thousands of exposures — is less than the current system. Still, Wu thinks the X-ray dose per exposure could be lowered enough that the method could come out ahead. Wu and the team acknowledged that image quality will also need to be improved, either with finer patterns or better computational methods to create the image.
Related on Eyes On Events: