For millions of people, slight camera movement during a photograph’s exposure leads to objectionable image blur ruining countless photographs every year; but assume this effect a problem of the past. SIGGRAPH 2006, the 33rd International Conference and Exhibition on Computer Graphics and Interactive Techniques held 30 July to 3 August 2006 in Boston, Massachusetts, reveals a ground-breaking technology that will change the photographer’s ability to capture clear images.

At SIGGRAPH 2006, leading researchers will present “Removing Camera Shake from a Single Photograph” (PDF here) that introduces a method to remove the effects of camera shake from seriously blurred images. This new method assumes a uniform camera blur over the image and negligible in-plane camera rotation. The result is that a photograph is transformed from blurry to clear.

Konquest Online will present each day previews of the applications that will be presented at SIGGRAPH with, let’s hope, some nice background info and relevant links to help you understand why the works presented at this conference is among the best in their category.

July 30th, 2006 | Technology | 1 comment

This week’s cover of Nature Materials showcases a new tool used to measure optical fields composed of optical fibers developped by MIT’s materials scientist Yoel Fink and physicist John Joannopoulos.

The fibers are made of a semiconducting glass core, lined along its full length by wires that act as positive and negative electrodes, and surrounded by a transparent polymer. When light hits the photosensitive core, an electrical current in the fiber changes, registering the hit.

A mesh of these fibers can then be used to identify the location of the light on a surface. The Nature Materials paper demonstrates that the fibers, in addition to locating a point of light, can be used to determine the direction from which a light beam comes and can also sense light from a scene to form an image.

The fibers are woven in a sphere, in a way that hits are registered when the light enters the sphere and again when it leaves the sphere. That way, it can deduce the direction of the incoming ray. With the information of location and direction, the researchers can reconstruct the scene in which the sphere is placed. However, because there’s no lens, which in a camera focuses light from a given plane onto a light detector, the grids receive a blurry image. To compensate for the lack of a lens, the researchers wrote algorithms that compare slight differences between the images recorded by the two fiber grids. These differences allow them to trace the light back to its source — and mathematically reconstruct an in-focus image. Because this “focusing” happens after the data has been recorded, it’s also possible to refocus on various objects in a scene after a picture has been taken.

But work is still needed to make the scene-imaging capabilities practical. For example, the resolution of the images is limited by the need to space the fibers within the grid far enough apart that the first grid does not distort the image received by the second. The grids themselves also need to be separated, which could make the current system difficult to incorporate into some applications, such as on the skin of a car, where keeping the grids at a distance wouldn’t be practical. But the researchers say work is currently being done that could overcome these limitations.

July 28th, 2006 | Physics | No comments