Digital Image Processing

A radio telescope with thousands of antennas is like a giant ear that hears the whole sky at once. But an ear cannot draw a picture: signals from every pair of antennas must be multiplied, and in an array of ten thousand elements there are already fifty million pairs. A classical correlator does exactly that, and its complexity grows as the square of the number of antennas.

You're looking for fireflies in a dark forest. Instead of photographing the whole forest every second and then poring over the images, you carry a small device that spots the flashes on its own and cuts out only a circle around each one. The rest, the blackness, isn't saved. That's exactly how the new sensor for scientific cameras works.

You're looking for fireflies in a dark forest. Instead of photographing the whole forest every second and then poring over the images, you carry a small device that spots the flashes on its own and cuts out only a circle around each one. The rest, the blackness, isn't saved. That's exactly how the new sensor for scientific cameras works.

A stenographer is recording a meeting, but halfway through the conversation she runs out of paper. She misses a few lines, then picks up where she left off. Later, she needs to reconstruct the full transcript without rewriting the entire conversation from scratch. That's exactly how the new method URA-SIM works for structured illumination microscopy (SIM).

A low-dose X-ray image looks like a view through a fogged-up window: outlines are there, but details drown in a hazy veil. Doctors need such images during radiation therapy — to aim the beam precisely at the tumor and spare healthy tissue. But saving on radiation comes at the cost of noise, streaks, and artifacts that obscure what matters.

The proton beam is like a sniper's bullet: it delivers most of its energy at a precisely set point, barely touching tissues on the way in. But over 4–6 weeks of head and neck tumor treatment, the target shifts: the patient loses weight, the tumor shrinks, posture changes. The sniper aimed at one spot, but the target has already moved.

A pebble on a streambed lies still as sunlight sparkles dance on the water's surface. In astronomy, the contrast is similar: a real exoplanet is the stone, while myriad speckles (patches of scattered starlight) are the ripples trembling in instruments.

In a three-dimensional scene, each point is like a candy whose color changes depending on the viewing angle. To describe this play of shades, dozens of numbers are stored for each point — a special mathematical function invented by Joseph Fourier. Memory fills up instantly.

A single-pixel camera works like an artist creating a picture through a stencil with many holes. In each frame, only a portion of the light, modulated by a mask, a spatial light modulator, reaches the single sensor, and then the computer reconstructs the full image by assembling this mosaic with computational algorithms.

The photo enhancement algorithm is like a navigator that step by step leads a blurry picture toward a sharp view. It removes noise and distortions, moving pixel by pixel along a familiar map. But if the image was taken in unfamiliar conditions (say, with unusual blur) the map doesn't match the terrain.

A drop of urine under a microscope isn't a flat puddle but a multi-layered volume. Like a stack of transparent pages: if you focus on one, the neighboring ones blur. The microscope's light only shows one depth layer sharply. To see all the cells, a lab technician had to take a series of snapshots, manually adjusting focus — tedious and slow.

The data from a single CT scan is comparable to several feature-length films. TCAM-Diff bypasses this bottleneck: the program compresses the 3D scene into three shadow projections — like a sculptor reconstructing a solid statue from silhouettes on three walls.

A traditional filter works like a long chain of people passing a message around: each waits for the previous one to finish. That's slow. Engineers applied a clever method — cyclic reduction. They rearrange data so that dependencies between samples become not long chains but small groups.

Trained on ordinary images, the DINOv3 model unexpectedly began to see faces as a spectrum: on its own, without hints, it isolates the eye, nose, and mouth regions. Like an invisible spectroscope, it breaks down a portrait into parts, just as a prism splits light into colors. Across 200 videos, feature labeling accuracy reached 83% (random guessing is 23%).

A cook who learns to cook from hand-drawn pictures risks ruining every dish. The same trap awaits computer vision: models need thousands of labeled images, but collecting them manually is expensive and tedious. Synthetic frames solve the problem, but only if they look real.

Assessing image quality is a job for a meticulous inspector with a toolbox. The IQA-T1 neural network acts exactly like that: it doesn't just stare at the picture, but applies either a spectroscope or a light meter.

A raw digital photo resembles a mosaic: each pixel holds only one base color — red, green, or blue. To compress such a file, it's split into brightness (a detailed black-and-white image) and color components.

Modern machine vision systems are gaining autonomy: engineers have managed to implement an image recognition algorithm directly on a miniature chip, freeing the camera from having to send data to a bulky computer.

Prostate MRI scans often look like photos taken through a dirty window: noise, blurred edges, curvature from field imperfections. Scientists checked what happens if you 'wash' that window: remove the excess, correct the distortions, and accurately align different types of scans.

Building a light chip is like a game where you have to guide a beam through a maze of tiny mirrors and beam splitters. Each element acts like a railroad switch: it either lets light pass straight through or directs it to an adjacent track. The goal is to produce a specific light pattern at the output while using as few 'switches' as possible. Usually, engineers spend years finding the shortest route for each task.

An electron microscope is a giant pipe organ: instead of pipes (electromagnetic lenses, instead of air) a stream of accelerated electrons. It once merely painted mute frescoes of the atomic world, condemning scientists to grueling decipherment.

Look into the night sky — before you is not a scattering of lights, but a multi-layered manuscript. Under the bright ink of stars and galaxies appears a faded, almost erased text.

On the screen is the result of a complex simulation: two planets collided, a shockwave ripples outward. A couple of clicks in an editor, and harsh shadows soften, noise smooths out, the background darkens. This kind of digital retouching (image editing to enhance appearance) has long been part of the scientist’s toolkit.

A lab robot counts molecules, and gets it wrong. One tube shows 0, another 12, a third 3, though in reality they all hold roughly the same amount. Such 'raw' numbers are common in biology: DNA sequencing, counting species in a water sample, gene activity. There's plenty of data, but it's sparse and more scattered than simple randomness would predict.

Generative neural networks that create images typically rely on digital randomness. But a team of researchers built a system where randomness is born not in code, but in a tiny optical cavity filled with an organic dye.

A single camera shot stores both a normal picture and a hidden map of how light rippled through the scene. Quantum entanglement makes this possible. Pairs of linked photons are generated: one visits the object, its twin takes a detour. Later, measuring the twin one way pulls out the brightness information.

An artist paints a portrait stroke by stroke, and then someone offers them a flash camera — one click, and it's done. That's roughly the shift from multi-step models to one-step models in generating images of galaxies. Typical neural networks for image synthesis work like a watercolorist: they apply a layer, blur it, add details.

A solar telescope is like an eye trying to spot an ant through a pot of boiling soup. The daytime sky dazzles, the air shimmers, and the image swims. To freeze it, you need adaptive optics — a system that predicts distortions and corrects them hundreds of times per second.

The vessels inside the skull resemble a mountain road's hairpin turns — sharp bends, narrowings, unexpected curves. When a traffic jam forms on such a serpentine, blood stops reaching the brain properly. To predict trouble, doctors need to see the entire flow: where it accelerates, where it swirls into eddies. The full picture comes from a special MRI, like slow-motion cinematography of fluid.

Any light measurement runs into a fundamental limit: photons hit the camera sensor unevenly. This natural ripple (shot noise) blurs fine details, like trying to pick out a quiet instrument in an orchestra through the noise of the hall.

A glass chip, threaded with 128 optical waveguides, has been turned into a kind of quantum pinball machine. The role of balls is played by photons — quanta of light, and instead of mechanical flippers, microheaters control the routes, instantly altering the optical paths.

A magnetic film just tens of atoms thick stores information in tiny magnetic vortices. But you can't look inside and see these vortices directly: the instrument only senses the weak magnetic field around the film. Reconstructing the magnetization pattern from such a field is like guessing the shape of an invisible relief from its shadow on a wall.

A typical computer shuttles electricity along copper traces, and every operation is a queue: first one thing, then another. Light works differently. In a single beam, you can mix dozens of signals without them interfering, like different voices in a choir, singing at once but each heard distinctly. The new FieldCore chip uses five such "voices" at once.

The X-ray machine moves around the breast in an arc, taking several images from different angles. From these flat frames, the computer assembles a 3D picture, like putting together a mosaic from pieces. But if the mosaic pieces are crumpled or smudged, the picture turns out blurry. In tomography, the role of "crumples" is played by noise and imperfections in the imaging.

A conventional lidar is like an artist who paints a landscape but sees only gray shadows. It measures distances to each point accurately, but can't tell whether it's looking at rock, wood, or plastic. To learn the composition, you need a spectrometer — a separate device that works slowly and can't keep up with a fast scanner. The new system takes a different approach.

A balloon with a dried drop of glue inflates unevenly: the glue stretches differently than the rubber. The cornea, the eye's transparent outer layer, behaves similarly after procedures that alter its elasticity. Engineers tested this on pig eyes.

An orchestra where each musician plays their own part but hears only the neighbor. One wrong note, and a wave of errors rolls through the hall. That's how Ising machines work — special devices for solving problems where you need to find the best option among many.

Physical systems can compute. If you build a landscape of hills and valleys so that the lowest point corresponds to the answer, the system will roll there by itself, like a ball. The problem is that the bottom is often flat, and the ball crawls around for a long time before settling. The authors propose nudging the ball to the right place before it even starts.

A laser melts metal powder layer by layer, and inside the boiling pool, vapor bubbles appear and collapse within milliseconds. Seeing this kitchen in action was previously impossible: the X-ray tomograph that takes 3D images is too slow. The sample has to rotate, and by the time it does, the melt has already changed a hundred times.

Fibers in composites, paper, or biological tissues are often hidden from view, but their orientation determines the strength and properties of the material. Previously, to see fiber directions, you had to take many images at different angles and assemble a three-dimensional picture, like slicing a loaf of bread into thin pieces. For thin samples, this approach doesn't work: you can't slice what isn't thick enough.