- Atlas
- Medicine & Health
- Imaging Physics
Imaging Physics
Imaging physics is the branch of medical physics that asks how a scanner can turn a physical interaction—X-rays passing through tissue, hydrogen nuclei wobbling in a magnetic field, sound waves echoing off boundaries—into a picture a clinician can read.
10 frameworks, from 1895 to today · Medical Imaging · Medicine & Health
- Radiography 1895-Present
- Dosimetry 1928-Present
- Signal Detection and Image Quality Theory 1940-Present
- Nuclear Medicine Physics 1950-Present
- Ultrasound Physics 1950-Present
- Tomographic reconstruction 1971-Present
- Physics of magnetic resonance imaging 1973-Present
- Digital Detector Physics 1980-Present
- Hybrid and Multimodality Imaging 1990-Present
- Computational and Model-Based Imaging 2000-Present
Overview
Imaging physics is the branch of medical physics that asks how a scanner can turn a physical interaction—X-rays passing through tissue, hydrogen nuclei wobbling in a magnetic field, sound waves echoing off boundaries—into a picture a clinician can read. Its practitioners are physicists and engineers who work alongside radiologists and radiographers. They design detectors, model how signals and noise propagate, calibrate machines, and set the limits on what an image can show and what dose a patient receives. What is at stake is both diagnostic truth and safety: an image that is too noisy or too distorted can hide a tumour or a fracture, and the radiation or energy used to make it can itself cause harm. The history of imaging physics is therefore a history of learning to measure the body precisely, and of learning to measure the measurement.
Shadows on Film: The First Half-Century
Medical imaging began with the shadow. Wilhelm Röntgen's discovery of X-rays in 1895 produced the first radiograph within weeks, and the physics of projection radiography was understood almost immediately in outline: X-rays are absorbed or scattered according to the density and atomic composition of the material they cross, so a detector on the far side records a two-dimensional shadow of everything in the beam's path. The early problems were practical—film sensitivity, exposure time, and the fact that a shadow collapses depth. Practitioners learned to take two views at right angles to recover position, and the field of radiology grew around the reading of these shadows.
The physics of the X-ray tube itself became a discipline. Physicists worked out how the spectrum of emitted X-rays depends on tube voltage and target material, and how filtration shapes the beam. By the mid-twentieth century, the "physics of radiology" was a recognized subject, taught to radiologists and radiographers, with textbooks covering beam quality, scatter, grid design, and film-screen response. Radiation protection grew alongside it: the recognition that ionizing radiation damages tissue led to the founding of the International Commission on Radiological Protection in 1928 and to the gradual codification of dose limits. The central concept—absorbed dose, the energy deposited per unit mass of tissue—and its operational measures became the vocabulary of the field1. Radiation dosimetry and protection was never a side topic; it was the conscience of imaging physics, and it shaped every subsequent technology.
The Image as a Signal: Detection Theory and Quality
A shadow is only useful if it can be distinguished from noise. In the mid-twentieth century, physicists began to treat the radiographic image as a signal-detection problem. The key insight, developed through the 1940s to 1960s, was that image quality is not a single number but a relationship between the contrast of a feature, the noise in the image, and the observer's task. The modulation transfer function (MTF) described how well a system preserves spatial detail; the noise power spectrum described the texture of the randomness; and the detective quantum efficiency (DQE) combined them into a measure of how efficiently a detector converts incoming X-ray quanta into usable signal. These ideas came from optics and communication theory and were adapted to radiology by physicists such as Kurt Rossmann and others working on image evaluation.
The practical consequence was that detector design could be judged against a task rather than by eye. This mattered when digital detectors arrived. Computed radiography, using photostimulable phosphor plates, appeared in the 1980s, and flat-panel detectors using amorphous silicon or selenium followed in the 1990s. Digital detector physics became its own subject: how X-rays are converted to charge, how that charge is read out, how pixels and fill factors affect resolution, and how the DQE of a digital system compares with film. The signal-detection framework gave the field a common language for comparing systems and for setting exposure levels that balance image quality against dose.
Slices and Mathematics: Tomography and Reconstruction
Projection radiography's fundamental limitation is superposition: everything along a ray is summed. The solution—computed tomography—was conceived by Godfrey Hounsfield and Allan Cormack independently, with Hounsfield's first clinical scanner operating in 1971. The physics problem was to measure X-ray attenuation along many paths through a slice and then reconstruct the distribution of attenuation from those line integrals. The mathematical tool was the Radon transform, known since 1917, and the reconstruction method that made it practical was filtered backprojection: backproject the measured profiles and apply a filter to correct the blurring. The result was a quantitative map of tissue density, calibrated in Hounsfield units, that removed the depth ambiguity of the shadow.
Tomographic image reconstruction became a central branch of imaging physics. It was not confined to X-rays: positron emission tomography (PET) and single-photon emission computed tomography (SPECT) reconstruct distributions of radioactive tracers from emitted gamma rays, and magnetic resonance imaging reconstructs from frequency-encoded signals. Each modality posed its own reconstruction problem, and the field developed a shared toolkit—Fourier methods, iterative algorithms, and later statistical and model-based approaches. Iterative reconstruction, which starts with a guess and refines it against the measured data, was known early but became practical as computing power grew; it allowed lower dose and better handling of noisy data. The reconstruction step is where physics, mathematics, and engineering meet, and it remains one of the most active parts of the field.
Nuclei in a Magnetic Field: MRI Physics
Magnetic resonance imaging rests on a different physical interaction. Atomic nuclei with spin, notably hydrogen protons abundant in water and fat, align in a strong magnetic field and precess at a frequency proportional to field strength—the Larmor frequency23. Radio-frequency pulses tip them out of alignment, and as they relax they emit signals that depend on the surrounding tissue. Paul Lauterbur in 1973 showed that magnetic field gradients could encode spatial position into the signal, and Peter Mansfield developed the echo-planar techniques that made fast imaging possible. The first human MRI images followed in the late 1970s.
MRI physics became a rich subfield because the signal is so malleable. The relaxation times T1 and T2 differ between tissues, and pulse sequences can be designed to weight the image toward one or the other24. Gradients encode position in frequency and phase; the Fourier relationship between the measured signal and the image is the basis of reconstruction. The physics problems are distinctive: field homogeneity, gradient performance, radio-frequency coil design, and the safety issues of strong fields and absorbed radio-frequency power. Functional MRI, diffusion imaging, and spectroscopy all grew from the same physics, each exploiting a different aspect of nuclear behavior. MRI is often described as the modality with the most physics per pixel, and its history is a long refinement of what can be encoded and how cleanly it can be decoded.
Sound and Tracers: Ultrasound and Nuclear Medicine
Two other modalities developed their own physics traditions. Pulse-echo ultrasound sends high-frequency sound into the body and listens for echoes from boundaries between tissues of different acoustic impedance2. The physics of attenuation, scattering, and beam formation determines resolution and penetration depth; the field developed array transducers, focusing, and later harmonic and Doppler techniques. Ultrasound's advantages—no ionizing radiation, real-time imaging, low cost—made it ubiquitous, and its physics is the study of how sound propagates and how echoes can be turned into a picture.
Nuclear medicine physics concerns tracers: radioactive substances injected into the body that emit gamma rays or positrons25. The physics problems are tracer kinetics—how a substance distributes and clears over time—and detection: collimators, scintillation crystals, and photomultiplier tubes that convert gamma rays into signals. PET exploits positron annihilation to detect pairs of gamma rays in coincidence, which localizes the emission event along a line. The field's history includes the development of radiopharmaceuticals, the physics of detector sensitivity and resolution, and the reconstruction of tracer distributions. It is a branch where physiology and physics are inseparable, because the signal is the tracer's behavior.
Combining Modalities and Computing the Image
As the individual modalities matured, the field turned to combination. Hybrid and multimodality imaging—PET/CT, SPECT/CT, and later PET/MRI—posed the problem of aligning two different physical measurements of the same anatomy. The physics challenges are registration, attenuation correction (using the CT to correct PET data), and the design of detectors that can work in both systems. Hybrid imaging became clinically important because it combines functional information with anatomical localization26.
The most recent turn is computational and model-based imaging. Instead of treating reconstruction as a fixed mathematical inversion, the field increasingly builds physical models of the imaging process into the reconstruction, and uses computation to solve problems that were previously intractable. Compressed sensing, which exploits sparsity to reconstruct from fewer measurements, reduced MRI scan times; model-based iterative reconstruction improved CT image quality at low dose; and machine learning entered the reconstruction pipeline. The physics questions remain: what is the forward model, what is the noise model, and what does the reconstruction assume about the object? These are the questions that connect the earliest shadow physics to the present.
The Present: Shared Frameworks and Open Questions
Imaging physics today is organized around the same problems that drove its history, now expressed in a common language. Image quality and signal detection theory provide the framework for evaluating any system; tomographic reconstruction and computational imaging provide the methods for turning measurements into pictures; dosimetry and protection set the safety envelope; and the physics of each modality—projection radiography, digital detectors, MRI, ultrasound, nuclear medicine—supplies the specific interactions. Hybrid systems and model-based methods are where the field is most active, because they combine physical measurement with computation in ways that earlier generations could not.
The open questions are the old ones in new form. How low can dose go while preserving diagnostic truth? How can reconstruction incorporate accurate physical models without becoming unstable? How should machine learning be validated when it changes the image in ways that are hard to characterize? And how do hybrid systems correct one modality's data using another's without introducing new errors? These are questions of measurement and inference, and they are the inheritance of a century of learning to see into the body without cutting it open.
References
- Absorbed dose, Wikipedia
- Medical imaging (Magnetic resonance imaging, Ultrasound, Nuclear medicine), Wikipedia
- Physics of magnetic resonance imaging, Wikipedia
- Magnetic resonance imaging (History), Wikipedia
- Nuclear medicine, Wikipedia
- PET–CT, Wikipedia
The frameworks of Imaging Physics
Each framework is a way of working the field has used: when it held, and the question it answers.
- 1895-Present
- 1928-Present
- 1940-Present
Signal Detection and Image Quality Theory
How well can a signal be distinguished from noise?
- 1950-Present
Nuclear Medicine Physics
How do radioactive tracers distribute and emit detectable signals?
- 1950-Present
- 1971-Present
Tomographic reconstruction
How can we reconstruct a slice from line integrals?
- 1973-Present
Physics of magnetic resonance imaging
How do nuclear spins encode spatial information?
- 1980-Present
Digital Detector Physics
How do digital detectors convert X-rays to signals?
- 1990-Present
Hybrid and Multimodality Imaging
How can two different physical measurements be aligned?
- 2000-Present
Computational and Model-Based Imaging
How can computation improve image reconstruction?
How the ideas connect
Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.
Dosimetry grew out of Radiography
Radiography's use of ionizing X-rays created the need to quantify tissue damage, leading to the founding of the ICRP in 1928 and the development of absorbed dose concepts as the operational vocabulary of radiation protection.
Carried by International Commission on Radiological Protection.
Tomographic reconstruction influenced by Radiography
Computed tomography extended projection radiography by solving its superposition problem: Hounsfield's scanner measured X-ray attenuation along many paths and reconstructed a slice, using the Radon transform and filtered backprojection.
Carried by Godfrey Hounsfield; Allan Cormack.
Digital Detector Physics influenced by Signal Detection and Image Quality Theory
The signal-detection framework, especially DQE and MTF, provided the language to compare digital detectors with film and to optimize exposure, guiding the development of computed radiography and flat-panel detectors.
Hybrid and Multimodality Imaging influenced by Nuclear Medicine Physics
PET/CT and SPECT/CT combined nuclear medicine detectors with X-ray CT, using the CT for attenuation correction and anatomical localization, directly building on nuclear medicine physics.
Computational and Model-Based Imaging grew out of Tomographic reconstruction
Iterative reconstruction, known early in tomography, became practical with growing computing power and evolved into statistical and model-based approaches that form the core of computational imaging.
Signal Detection and Image Quality Theory influenced by Radiography
The radiographic image was reconceived as a signal-detection problem, with concepts like the modulation transfer function and detective quantum efficiency adapted from optics and communication theory to evaluate film-screen systems.
Carried by Kurt Rossmann.
Physics of magnetic resonance imaging influenced by Tomographic reconstruction
MRI reconstruction exploited the Fourier relationship between measured signals and images, a mathematical approach central to tomographic reconstruction, as shown by Lauterbur's use of gradients and later Fourier imaging.
Carried by Paul Lauterbur.
Hybrid and Multimodality Imaging influenced by Physics of magnetic resonance imaging
PET/MRI hybrid systems required combining MRI physics with nuclear medicine detection, posing challenges of compatibility and alignment that drove multimodality integration.
Computational and Model-Based Imaging influenced by Digital Detector Physics
Digital detector physics provided the precise noise and signal models that model-based reconstruction algorithms use to improve image quality at lower dose.
Beyond the field
Where Imaging Physics's ideas came from in other fields, and what they led to there: the strongest 12 of 66 links.
Radiography influenced Uroradiology and Diagnostic Imaging in Urology
Röntgen's discovery of X-rays and the rapid development of radiography provided the physical basis for visualizing the urinary tract, beginning with plain films of calculi and later contrast studies.
Carried by Wilhelm Conrad Röntgen.
Radiography led to Fluoroscopy in Medical Imaging
Fluoroscopy grew directly out of radiography within a year of Röntgen's discovery: instead of a photographic plate, a fluorescent screen made the X-ray shadow visible in real time. It is a branch of the same radiographic technique.
Carried by Wilhelm Conrad Röntgen; Thomas Edison.
Radiography influenced Oral and maxillofacial radiology in Dentistry
Oral and maxillofacial radiology is the application of radiography to the teeth and jaws, taking Roentgen's discovery and its dental adaptation directly. The specialty is unthinkable without radiographic technique.
Carried by Wilhelm Conrad Röntgen; C. Edmund Kells.
Dosimetry grew out of Empirical radiotherapy in Radiation Oncology
Dosimetry grew directly out of empirical radiotherapy: early practitioners dosed by skin reaction and exposure time, and the need to standardise the unit of quantity, culminating in the röntgen and the ICRU in 1928, produced the measurement science. Dosimetry is the refinement of empirical radiotherapy's trial-and-error dosing into a quantitative discipline.
Carried by International Commission on Radiation Units and Measurements.
Nuclear Medicine Physics influenced by Tracer Principle in Molecular Imaging
The tracer principle, introduced by Hevesy in 1913 and developed through the 1920s, is the conceptual foundation of nuclear medicine: a radioactive substance traces a physiological pathway without perturbing it. Nuclear medicine physics formalised the detection and quantification of such tracers.
Carried by George de Hevesy.
Nuclear Medicine Physics influenced by Radioactivity and Nuclear Spectroscopy in Nuclear Physics
Radioactivity and nuclear spectroscopy provided the decay schemes, gamma energies and half-lives that nuclear medicine physics uses to choose and detect tracers, and the scintillation counting methods that became the gamma camera. The whole physical basis of nuclear medicine derives from this earlier nuclear physics.
Carried by Ernest Rutherford; Hal Anger.
Ultrasound Physics influenced by Ultrasound in Acoustics
Ultrasound physics in acoustics, from Langevin's wartime sonar onward, supplied the piezoelectric transducers, pulse-echo ranging and attenuation theory that medical ultrasound physics applied to tissue. The medical field is a direct application of this earlier acoustics.
Carried by Paul Langevin.
Tomographic reconstruction influenced by Inverse problem in Computational Imaging
The theory of inverse problems, especially the Radon transform and its inversion, is the mathematical core of tomographic reconstruction, and Cormack and Hounsfield drew on it directly. The later framework is the application of this earlier mathematics to medical imaging.
Carried by Johann Radon; Allan M. Cormack; Godfrey Hounsfield.
Tomographic reconstruction influenced Deep-learning reconstruction and analysis in Quantitative Imaging
Deep-learning reconstruction builds on classical tomographic reconstruction, using its forward models and filtered-backprojection or iterative formulations as the structure networks learn to replace or accelerate. The mathematical problem it solves is the tomographic one.
Tomographic reconstruction influenced Seismic Tomography in Geophysics
Seismic tomography adapted the mathematical inversion methods of medical tomographic reconstruction to travel-time data from earthquakes. The reconstruction algorithms were transferred from imaging physics to seismology.
Tomographic reconstruction influenced Positron Emission Tomography in Neuroimaging
PET image formation rests on tomographic reconstruction algorithms, the filtered back-projection and iterative methods developed for X-ray and emission tomography.
Carried by Michael Phelps.
Tomographic reconstruction led to Model-Based Reconstruction in Image Reconstruction
Model-based reconstruction is a branch of tomographic reconstruction that replaces analytic inversion with an explicit system model and iterative statistical estimation. It grew out of the tomography literature and its forward models.
Carried by Jeffrey Fessler.

Imaging Physics is a planet in Nousfarer
Its city has a tower for each of these 10 frameworks. Land, walk in, and climb one concept at a time.
Land on Imaging PhysicsRelated fields
Fields its ideas are linked to
- Molecular Imaging 13 links
- Medical Imaging 9 links
- Quantitative Imaging 6 links
- Computational Imaging 4 links
- Image Reconstruction 4 links
- Neuroimaging 4 links
- Radiation Oncology 4 links
- Nuclear Physics 3 links
- Inverse Problems 2 links
- Photography Theory 1 link
- Biomedical Signal Processing 1 link
- Spectroscopy 1 link
More in Medical Imaging
- Diagnostic Radiology
- Image Reconstruction
- Imaging Informatics
- Interventional Radiology
- Medical Image Analysis
- Medical Imaging
- Molecular Imaging
- Nuclear Medicine
- Quantitative Imaging
See where Imaging Physics sits on the Map of Human Knowledge