Bringing it all into focus
UC Berkeley-led researchers developed a computational microscope combining 48 sensors and advanced algorithms to achieve 3-micron resolution across 5 square centimeters at 120 frames per second, surpassing traditional trade-offs in speed, field of view, and resolution.
Optical engineers have historically struggled to balance speed, field of view, and resolution in microscopes, as improving one often sacrifices another. A UC Berkeley-led team addressed this by designing a computational microscope that integrates 48 camera sensors with a diffractive optical element and optimization algorithm. The system captures micron-scale resolution across multi-centimeter areas at 25.2 billion pixels per second, enabling unprecedented imaging capabilities. Professor Laura Waller, the study’s principal investigator, noted the breakthrough expands possibilities for imaging dynamic biological samples with high fidelity.
The microscope overcomes a long-standing limitation in spatiotemporal throughput, where traditional systems cannot simultaneously achieve high resolution, wide field of view, and rapid frame rates. Researchers engineered a phase mask to redirect light falling between sensors onto the sensor array, while a computational algorithm reconstructed missing data. The design eliminates the need for manual calibration, reducing time and effort. The team demonstrated the system’s versatility by imaging static samples with higher resolution than traditional methods and tracking freely moving C. elegans nematodes at 120 frames per second.
Lead author Kevin C. Zhou, formerly a UC Berkeley postdoctoral researcher, highlighted the microscope’s ability to track individual worms and perform functional imaging of their feeding behavior, such as pharyngeal pumping. The system’s simultaneous high-resolution, wide-field, and high-speed capabilities were critical for capturing rapid biological processes. Co-author Chaoying Gu emphasized the calibration-free approach as a key advancement for scaling future imaging systems. The research team included collaborators from UCSF, University of Utah, UC San Diego, and Duke University.
Funded by the Office of Naval Research, Air Force Office of Scientific Research, National Institutes of Health, and Japan Society for Promotion of Science, the study underscores the potential of computational imaging. Waller, a Chan Zuckerberg Biohub SF investigator, noted the work advances large-scale video microscopy, demonstrating how joint hardware-software design can overcome traditional imaging constraints. The findings open new avenues for monitoring dynamic biological systems with unprecedented detail and speed.