Large-scale compressive microscopy via diffractive multiplexing across a sensor array - nature.com

What the report says

Nature.com reported on a computational microscopy approach designed to ease a long-standing constraint in optical imaging: the difficulty of increasing resolution, field of view and speed at the same time. The researchers describe a microscope that combines a 48-sensor array, a diffractive optical element and compressive sensing to produce gigapixel-scale images in a single snapshot.

The system treats the sensors as one large “super sensor,” even though physical gaps between the sensors mean much of the image plane is not directly recorded. A diffractive mask placed in the Fourier plane creates a multi-spot point spread function, spreading information so that areas falling into the gaps can still be inferred. Reconstruction software then solves a large inverse problem, using assumptions about sparsity in the sample to recover the missing image data.

According to the Nature article, the microscope achieved about 3-micrometre resolution over more than 5.2 square centimetres at up to 120 frames per second, with an effective throughput of 25.2 billion pixels per second. The sensor array captures 0.617 gigapixels per full snapshot, though the reported high-speed demonstrations used 4-by-4 pixel binning to reach 120 fps.

The authors demonstrated both structural dark-field imaging and fluorescence imaging of calcium activity in dozens of freely moving Caenorhabditis elegans. If further validated and adapted, the technique could be useful for studying dynamic biological systems over centimetre-scale areas, where conventional microscopes often must trade speed for coverage or detail.