Novel Method Enhances Accuracy of Imaging Systems

The enhanced method for illustrating the position of objects by imaging systems is the result of novel study by researchers at the University of Chicago. The study offers a mechanism – known as single pixel interior filling function, or SPIFF to correct and identify systematic errors in image and data analysis employed in numerous areas of engineering and science.

“Anyone working with the imaging data on small objects – or objects that appear small – who intends to determine and track their positions in space and time will advantage from the single-pixel interior filling function method,” says co-principal investigator Norbert Scherer, an UChicago chemistry lecturer.

Scientists across the sciences employing imaging to learn about objects on scales varying from the highly small, like nanometres, to the very big, such as astrophysical scales. Their work often comprises tracking the movement of such objects to learn about their properties and behaviour. Numerous imaging systems and image-based detectors are constituted of pixels, like with a mega-pixel cell phone.

So-called particle tracking enables scientists to determine the position of an object down to a single pixel and even explore sub-pixel localization to better than 1-10th of a pixel accuracy. With an optical microscope’s resolution of about 250 nanometres and an efficient pixel size of about 80 nanometres, particle tracking enable researchers to locate the centre or a location of an object to within a few nanometres, provided enough photons are estimated.

But such sub-pixel resolution depends on algorithms to calculate the position of objects and their trajectories. Employing such algorithms often results in errors of accuracy and precision due to factors like nearby or overlapping objects in the background noise and image.

SPIFF can rectify the errors with little added computational costs according to Scherer. “Till this work, there were no simple methods to determine if the tracking and sub-pixel localization was precise and to correct the error if it was not,” he says. “Analysing an image to avail a rough estimate of an object’s position is not too intricate, but making optimal use of all the data in an image to avail the finest possible tracking data can be truly challenging,” says David Grier, lecturer of physics at New York University, who was not involved in the study. “Given how extensively image-based particle tracking has penetrated physics, biology, chemistry and numerous engineering disciplines, this way should be exceedingly adopted.”

Sub-pixel information analysis can be biased by subtle features of the image-creation process, according to Grier, and such biases can shift a trajectory’s apparent placement by as much as half of a pixel relative to its true position.

The scientists have now applied this method to numerous other datasets, comprising nanoscale features of cells, single molecules and metallic nanoparticles, confirms Scherer, adding that the SPIFF method is applicable to all tracking algorithms