A new technique for super-resolution digital microscopy
January 27, 2016
Researchers from the California NanoSystems Institute at UCLA have created a new technique using lens-free holograms that greatly enhances digital microscopy images, which are sometimes blurry and pixelated.
The new technique, called “wavelength scanning pixel super-resolution,” uses a device that captures a stack of digital images of the same specimen, each with a slightly different wavelength of light. Then, researchers apply a newly devised algorithm that divides the pixels in each captured image into a number of smaller pixels, resulting in a much higher-resolution digital image of the specimen.
The research team was led by Aydogan Ozcan, Chancellor’s Professor of Electrical Engineering and Bioengineering at the UCLA Henry Samueli School of Engineering and Applied Science. The study appears in an open-access paper in the journal Light: Science and Applications, published by the Nature Publishing Group.
Faster, more-accessible diagnosis
“These results mean we can see and inspect large samples with finer details at the sub-micron [nanoscale] level,” Ozcan said. “We have applied this method to lens-based conventional microscopes, as well as our lensless on-chip microscopy systems that create microscopic images using holograms, and it works across all these platforms.”
The benefits of this new method are wide-ranging, but especially significant in pathology, where rapid microscopic imaging of large numbers of tissue or blood cells is key to diagnosing diseases such as cancer. The specimens used in the study were blood samples, used to screen for various diseases, and Papanicolaou tests, which are used to screen for cervical cancer.
Ozcan said that wavelength scanning super-resolution works on both colorless and dye-stained samples. The entire apparatus fits on a desktop, so its size and convenience could be of great benefit to doctors and scientists using microscopes in resource-limited settings such as clinics in developing countries.
The research was supported by the Presidential Early Career Award for Scientists and Engineers, the Army Research Office, the National Science Foundation, the Office of Naval Research and the Howard Hughes Medical Institute.
Abstract of Pixel super-resolution using wavelength scanning
Undersampling and pixelation affect a number of imaging systems, limiting the resolution of the acquired images, which becomes particularly significant for wide-field microscopy applications. Various super-resolution techniques have been implemented to mitigate this resolution loss by utilizing sub-pixel displacements in the imaging system, achieved, for example, by shifting the illumination source, the sensor-array and/or the sample, followed by digital synthesis of a smaller effective pixel by merging these subpixel-shifted low-resolution images. Herein, we introduce a new pixel super-resolution method that is based on wavelength scanning and demonstrate that as an alternative to physical shifting/displacements, wavelength diversity can be used to boost the resolution of a wide-field imaging system and significantly increase its space-bandwidth product. We confirmed the effectiveness of this new technique by improving the resolution of lens-free as well as lens-based microscopy systems and developed an iterative algorithm to generate high-resolution reconstructions of a specimen using undersampled diffraction patterns recorded at a few wavelengths covering a narrow spectrum (10-30 nm). When combined with a synthetic-aperture-based diffraction imaging technique, this wavelength-scanning super-resolution approach can achieve a half-pitch resolution of 250 nm, corresponding to a numerical aperture of approximately 1.0, across a large field of view (>20 mm2 ). We also demonstrated the effectiveness of this approach by imaging various biological samples, including blood and Papanicolaou smears. Compared with displacement-based super-resolution techniques, wavelength scanning brings uniform resolution improvement in all directions across a sensor array and requires significantly fewer measurements. This technique would broadly benefit wide-field imaging applications that demand larger space-bandwidth products.