Russian Scientists Create an Ultra-Precise Digital Holography “Scanner”
Researchers in Russia say they have dramatically improved digital holography, boosting image reconstruction accuracy by up to 200 times by filtering out laser noise without losing fine detail.
Russian scientists have developed a technology that significantly increases the accuracy of image recognition in digital holography—a visualization method used to reconstruct 2D and 3D objects and widely applied in medicine, industry, and information systems. The work was carried out by researchers at National Research Nuclear University MEPhI.
One of the key obstacles in digital holography is visual “noise.” When holograms are recorded using lasers, random speckle patterns—often described as grain—appear in images. This noise interferes with accurate object recognition and makes it harder to reproduce fine structural details.
To address this, MEPhI scientists proposed analyzing not a single image, but an entire set of holographic images using specialized software. As the university explained, speckle noise appears randomly and differently in each image, while the true boundaries of the object remain constant.
By comparing a series of frames, the software can determine which visual elements are random artifacts and which belong to the object itself. The algorithm removes the noise and preserves only the correct image, increasing the accuracy of digital holograms by up to 200 times.
Built for Medicine and Industry
According to the researchers, the method reproduces fine details 50 percent more accurately than existing software solutions.
The technology could be used in medicine, where extremely precise imaging is critical; in industrial applications such as evaluating microchip quality or detecting microscopic defects; and in advanced microscopy systems. The method may also improve hologram quality in digital data encoding.
Together, the results point to a broader shift in imaging science: instead of simply smoothing out imperfections, researchers are learning to exploit statistical patterns in noise—turning a long-standing limitation of laser imaging into a source of dramatically sharper visual data.
