A new computational microscopy technique provides a more direct route to sharp images

For hundreds of years, the clarity and magnification of microscopes was ultimately limited by the physical properties of their optical lenses. Microscope manufacturers have pushed these boundaries by producing ever more complicated and expensive sets of lenses. Still, scientists had to decide between high resolution and a small field of view on the one hand, or low resolution and a large field of view on the other.

In 2013, a team of Caltech engineers introduced a microscopic technique called FPM (for Fourier Ptychographic Microscopy). This technology marked the advent of computational microscopy, the use of techniques that combine the imaging of conventional microscopes with computer algorithms that process the detected information in new ways to produce deeper and sharper images covering larger areas. FPM has since been widely adopted for its ability to acquire high-resolution images of samples while maintaining a large field of view using relatively inexpensive equipment.

Now, the same lab has developed a new method that can outperform FPM in its ability to acquire images without blurring or distortion, even with fewer measurements. The new technique, described in an article that appeared in the magazine The nature of communicationcould lead to advances in such areas as biomedical imaging, digital pathology and drug screening.

The new method, called APIC (for Angular Ptychographic Imaging with Closed-form Method), has all the advantages of FPM without what could be described as its biggest weakness—namely, that the FPM algorithm relies on running on one or more best guesses and then gradually adjust a little to arrive at an “optimal” solution that may not always be true to the original image.

Led by Changhuei Yang, the Thomas G. Myers Professor of Electrical, Bioengineering, and Medical Engineering and a researcher at the Heritage Medical Research Institute, the Caltech team realized that it was possible to eliminate this iterative nature of the algorithm.

Instead of relying on trial and error to find a solution, APIC solves a linear equation that details the aberrations or distortions caused by the microscope’s optical system. Once the aberrations are known, the system can correct them, essentially behaving as if it were ideal, providing clear images covering large fields of view.

“We get to solve a complex high-resolution field in a closed form because we now have a deeper understanding of what the microscope captures, what we already know, and what we really need to find out, so I don’t need any iteration,” says Ruizhi Cao, co-author of the paper, a former graduate student at Yang’s lab and now a postdoctoral fellow at UC Berkeley. “That way we can basically guarantee that we’re seeing the true final details of the sample.”

As with FPM, the new method measures not only the intensity of light seen by the microscope, but also an important property of light called “phase,” which is related to the distance the light travels. This property cannot be detected by the human eye, but it contains information that is very useful in terms of aberration correction.

It was in solving the information for this phase that FPM relied on trial and error, explains Cheng Shen, a co-author of the APIC paper who also completed work in Yang’s lab and is now a computer vision algorithm engineer. at Apple.

“We’ve shown that our method gives you an analytical solution and in a much more straightforward way. It’s faster, more accurate, and uses some deep insights into the optical system,” says Shen.

In addition to eliminating the iterative nature of the phase-solving algorithm, the new technique also allows researchers to collect clear images over a large field of view without repeatedly refocusing the microscope. For FPM, if the sample height varied by even a few tens of microns from one section to another, the person using the microscope would have to refocus for the algorithm to work.

As these computational microscopy techniques often involve stitching together more than 100 lower-resolution images to stitch together a larger field of view, this means that APIC can speed up the process much more quickly and avoid the potential introduction of human error in many steps.

“We developed a framework to correct aberrations as well as improve resolution,” Cao says. “These two options can potentially benefit a wider range of imaging systems.”

Yang says the development of the APIC is central to the broader scope of work his lab is currently working on to optimize image data input for artificial intelligence (AI) applications.

“Recently, my lab showed that artificial intelligence can outperform expert pathologists in predicting metastatic progression from simple histopathology specimens from lung cancer patients,” says Yang. “This predictive ability is exquisitely dependent on obtaining uniformly focused, high-quality microscopic images, which APIC is well suited for.”