CT Ring Artifacts: Causes, QC, and Correction

CT ring artifacts are concentric circles in the reconstructed image caused by one or more detector channels that consistently over- or under-estimate attenuation. Because a third-generation scanner's detector rotates with the tube, a faulty channel samples a ray at a fixed distance from the center of rotation on every view, so its error reinforces into a ring instead of averaging away. The fix is a combination of detector calibration and, where needed, sinogram-domain or iterative correction.¹²

Ring artifacts are one of the most recognizable scanner-based artifacts in CT, and unlike patient-based artifacts they point directly at the hardware: a detector element or module that has drifted out of calibration or failed. A medical physicist who understands how a channel error becomes a ring can localize the fault, judge its clinical significance, and decide whether the answer is recalibration, service, or a software correction.¹⁹

This guide explains the physics of ring formation, why the center of rotation is the most sensitive location, how rings corrupt CT numbers, how to detect them in quality control, and the correction strategies available — from routine air calibration to modern iterative and sinogram-domain methods.

Introduction

A ring artifact is a scanner-based artifact: it originates in imperfect detector behavior, not in the patient. Barrett and Keat, in their widely used review of CT artifacts, group artifacts into physics-based, patient-based, scanner-based, and reconstruction-based categories, and place ring artifacts squarely in the scanner-based group arising from imperfections in scanner function.¹

That classification matters because it tells you where to look. A beam-hardening streak is a physics problem you manage with filtration and correction algorithms; a motion blur is a patient problem you manage with positioning and faster acquisition. A ring is a detector problem. It tells the physicist that a specific channel — at a specific distance from isocenter — is not matching its neighbors.¹⁹

Ring artifacts became more prominent, not less, as CT pushed toward higher spatial resolution, thinner detector elements, and photon-counting detectors, because finer channels read out without the smoothing of coarse binning and any per-channel inconsistency is more visible.²⁸ Understanding the geometry of ring formation is therefore as relevant on a current scanner as it was on early systems. DRPS evaluates these artifacts as part of CT physics testing and medical physics consulting for imaging facilities across Florida, Maryland, Virginia, Washington DC, California, and Nevada.

Topic Explanation

What is a ring artifact?

A ring artifact is a full or partial circle, centered on the image isocenter, produced when a detector channel's response differs systematically from its neighbors. In the projection data (the sinogram), that channel contributes a vertical stripe — a constant error at a fixed detector position across all projection angles. When the sinogram is reconstructed by filtered backprojection, a stripe at a fixed radial detector position maps to a circle in the image.²¹⁰

A few terms are used throughout this guide:

• Detector channel (element). One readout cell in the detector array. A third-generation scanner has hundreds to thousands of channels across the fan.
• Center of rotation (isocenter). The axis about which the tube and detector rotate; it projects to the center of the reconstructed image.
• Sinogram. The raw projection data arranged as detector position versus projection angle. A point object traces a sinusoid; a single faulty channel traces a straight vertical stripe.¹⁰
• Gain and offset calibration. The per-channel correction (air calibration) that normalizes each channel's response so a uniform input produces a uniform output.

Why a stripe becomes a ring

In a third-generation geometry the detector rotates with the tube, so channel number is tied to a fixed geometric relationship with the rotation axis. A given channel always measures the ray whose perpendicular distance from isocenter — its impact parameter — is the same on every view. If that channel reads high or low, the error is deposited along every ray tangent to a circle of that radius, and backprojection reinforces the error on that circle. The result is a ring.¹²¹⁰

This is the key insight for localization: the radius of the ring tells you which channel is at fault, and the completeness of the ring (full circle versus arc) tells you whether the channel is bad over the full rotation or only part of it.²⁶ A fourth-generation design with a stationary detector ring breaks the fixed channel-to-radius mapping and is far less prone to classic rings, but essentially all modern clinical multidetector scanners are third-generation, so ring control rests on calibration and correction.¹

Key Technical Principles

The geometry: from sinogram stripe to image ring

Consider a rebinned, parallel-beam description where the signed sinogram coordinate $s$ is the distance of a ray from the isocenter and $\theta$ is the projection angle. A point object at polar position $(r,\varphi )$ traces a sinusoid in the sinogram:

A single miscalibrated channel contributes an error confined to a fixed detector position, $s=s_0$, for all angles $\theta$. The set of image points whose sinusoids pass through $s=s_0$ at every angle is the circle

centered on the isocenter. That is why a constant per-channel error in the sinogram appears as a ring of radius equal to the channel's impact parameter.²¹⁰ The impact parameter itself is set by how far the channel sits from the central ray:

where $n$ is the channel index, $n_c$ is the channel that views the isocenter, and ${\Delta }_{\text{iso}}$ is the detector element spacing projected to the isocenter.

Worked example: ring radius

Suppose a scanner has an effective detector element spacing at isocenter of ${\Delta }_{\text{iso}}=0.6\text{mm}$, and the drifting channel sits $100$ channels away from the central channel. The ring radius is:

so the ring appears at a radius of 6 cm from the image center. A channel adjacent to the central ray ($n\approx n_c$) produces a ring of near-zero radius — a dense dot at the center of rotation. This is the physical reason the central spot is both common and clinically treacherous: it lands on the anatomy most often at the center of the field of view.¹²

Worked example: CT-number error

A ring is not only a cosmetic circle; it carries a quantitative bias. If a channel's gain error produces an apparent change in the local linear attenuation coefficient $\Delta \mu$ along the ring, the Hounsfield-unit deviation is approximately

Taking ${\mu }_{\text{water}}\approx 0.019{\text{mm}}^{-1}$ at a typical CT effective energy, a small gain drift that shifts the apparent attenuation by $\Delta \mu =0.00019{\text{mm}}^{-1}$ gives:

A 10 HU ring is easily visible on a narrow soft-tissue window and is large enough to matter for quantitative CT — bone densitometry, coronary calcium scoring, and tissue characterization all depend on HU accuracy of a few units. This is why ring control is inseparable from CT number and HU calibration QC.¹⁹

Partial rings, bands, and center spots

Not every detector fault produces a complete circle. A channel that fails only over part of the gantry rotation, or an element that is intermittently noisy, produces a partial ring or arc; a group of adjacent bad channels produces a band of concentric rings; and the single worst case, the central channel, produces a dense spot at isocenter.¹²⁶ Recognizing these variants prevents misclassification — a faint arc is still a detector problem, not anatomy.

Comparison: ring artifact correction strategies

Prell and colleagues showed that correcting in Cartesian coordinates can create new artifacts and under-correct rings close to the center of rotation, and that a transform to polar coordinates — where a ring becomes a straight line that is easy to smooth — is a necessary step for efficient correction.² Later work built on this with total-variation and iterative formulations that suppress strong rings while preserving spatial resolution, and with learning-based and detector-shifting approaches aimed at simultaneous ring removal and noise reduction in low-dose data.⁵⁶⁷⁸

Clinical Impact

Ring artifacts can both mimic and mask disease, and they degrade quantitative CT. A central spot or inner ring can overlie the posterior fossa, brainstem, spinal canal, or a small central lesion, where it may be mistaken for a calcification, a hypodense lesion, or an enhancing focus — or where it hides a true finding. Because the artifact is fixed to isocenter, it tends to recur in the same anatomic region across patients until the underlying detector problem is fixed.¹

Quantitatively, the HU bias along a ring undermines any task that depends on absolute attenuation. Coronary calcium scoring thresholds, bone mineral density, dual-energy material decomposition, and radiomic texture features are all sensitive to systematic HU shifts of the magnitude a ring can produce.¹⁹ A ring passing through a region of interest can change a reported number without changing the obvious appearance of the scan, which is the most dangerous failure mode because it is quiet.

Finally, rings interact with reconstruction. Iterative and deep-learning reconstruction can suppress some ring energy, but they can also propagate or transform a strong ring if the underlying projection error is severe, so a visible ring should prompt a calibration or service check rather than reliance on reconstruction alone.⁸ For the broader catalog of scanner- and physics-based artifacts, see our guide to CT image artifacts and their correction.

Practical Optimization Tips

1. Keep detector calibration current

Most rings never reach a clinical image because routine air (gain and offset) calibration normalizes each channel. Perform calibrations on the manufacturer's recommended schedule and after any service, and do not defer a calibration flagged by the scanner. A ring that appears on patient images despite a recent calibration suggests drift or a failing element that calibration alone will not hold.⁹

2. Make ring detection part of QC

Scan a uniform water or CT-number phantom and inspect a narrow window for faint rings; evaluate uniformity and noise across the field of view rather than only at the center. The AAPM Task Group 233 framework for CT performance evaluation provides task-based image-quality methods — noise, uniformity, and resolution — within which a developing ring shows up as a localized uniformity or noise disturbance before it becomes clinically obvious.⁹ Tie this to CT acceptance testing and commissioning and routine QC.

3. Localize before you correct

Use the ring radius to identify the offending channel region and the ring's completeness to judge whether the fault is continuous or intermittent. A reproducible full ring at a fixed radius points to a single drifted or dead channel; a band points to a module.²⁶ This localization tells service engineers where to look and tells the physicist whether software correction is a stopgap or an acceptable solution.

4. Use software correction as a complement, not a cover-up

Sinogram-domain and iterative ring corrections are powerful and, done well, preserve resolution — but they treat the symptom. If a channel has failed, correct the hardware; use software correction to recover studies already acquired and to clean up residual low-amplitude rings.²⁵⁸

Common pitfalls to avoid

• Blaming the patient. A concentric circle centered on isocenter is not motion or metal; it is the detector.¹
• Over-filtering. Aggressive Cartesian smoothing can blur real structure and introduce secondary artifacts; polar-domain methods are generally safer near the center.²
• Ignoring the central spot. The lowest-radius ring is the easiest to dismiss and the most likely to overlie critical anatomy.¹
• Trusting reconstruction to hide it. Iterative or deep-learning reconstruction may reduce a ring but should not substitute for fixing a failed detector.⁸
• Missing the quantitative bias. A ring through a region of interest can shift HU without an obvious visual cue.⁹

Regulatory Considerations

CT performance, including image uniformity and artifact behavior, is governed by equipment standards, accreditation requirements, and state radiation-machine rules. Ring artifacts are not called out by name in most regulations, but they fall under the general expectations for image quality, uniformity, and CT-number accuracy that acceptance testing and periodic QC are designed to verify.⁹

• AAPM Report No. 233 (Task Group 233) provides contemporary methods for CT system performance evaluation, including noise, uniformity, and task-based image quality that bracket ring detection.⁹
• ACR CT accreditation requires a qualified medical physicist's annual survey and routine QC that include uniformity and artifact assessment; a visible ring would be a finding to resolve.
• FDA 21 CFR 1020.33 sets federal performance standards for CT equipment, and state radiation-control programs regulate the installation and use of CT as a radiation-producing machine. In Florida, CT systems are administered under the state's radiation-machine rules; DRPS also serves Maryland, Virginia, Washington DC, California, Nevada, New York, Pennsylvania, New Jersey, and Delaware, where state radiation-control authorities impose parallel acceptance and QC expectations. Always confirm the requirements with the authority having jurisdiction.

A documented QC program that detects, localizes, and resolves ring artifacts — and distinguishes a calibration fix from a service need — is what makes a CT quality program defensible during accreditation review or inspection. DRPS supports this through CT physics testing and accreditation support.

Frequently Asked Questions (FAQs)

What causes ring artifacts in CT?

Ring artifacts are caused by one or more detector channels that are miscalibrated, drifting, or defective. In a third-generation CT scanner each channel views a ray at a fixed distance from the center of rotation on every projection, so a consistent gain or offset error in that channel reinforces into a concentric ring in the reconstructed image rather than averaging out.¹²

Why is the ring at the center of the image the most important?

A channel that views the ray through the isocenter maps to a ring of zero radius — a dot or dense spot at the center of rotation. Because the center of an axial image often contains clinically critical anatomy such as the posterior fossa, spinal canal, or small central lesions, a central ring or spot can mimic or obscure pathology and is the hardest to dismiss.¹

How do you tell a ring artifact from real anatomy?

A ring artifact is a smooth, complete or partial circle centered on the image isocenter, it appears at the same radius on every slice acquired with the affected detector, and it does not follow anatomic boundaries. Rotating the patient or rescanning does not move it relative to isocenter, which distinguishes it from a true structure.¹

Can a ring artifact change CT numbers?

Yes. A detector gain or offset error shifts the reconstructed attenuation along the ring, so CT numbers on and near the ring can be biased by several Hounsfield units or more. This matters for quantitative tasks such as bone density, coronary calcium scoring, and tissue characterization, where a small systematic HU error changes the result.¹⁹

How are ring artifacts corrected?

The first defense is detector calibration: air (gain and offset) calibration and, when needed, factory recalibration or replacement of a failed detector module. Residual rings are reduced with software that detects and smooths the corresponding stripes in the sinogram, usually after a transform to polar coordinates, or with iterative reconstruction that penalizes ring structure.²⁵⁹

Which CT scanners are prone to ring artifacts?

Third-generation scanners, in which the detector array rotates with the tube so each channel always samples the same radial position, are inherently susceptible. Fourth-generation designs with a stationary detector ring and cone-beam or flat-panel systems have different sensitivities, but modern clinical scanners are overwhelmingly third-generation, so ring control depends on detector calibration and correction software.¹²

Key Takeaways

• A ring is a detector problem. It is a scanner-based artifact from a miscalibrated, drifting, or dead channel — not a patient or physics artifact.¹
• A sinogram stripe becomes an image ring. A fixed per-channel error across all views maps to a circle of radius equal to the channel's impact parameter; the central channel maps to a spot at isocenter.²¹⁰
• Rings carry a quantitative bias. Even a small gain drift can shift CT numbers by ~10 HU along the ring, affecting quantitative CT.¹⁹
• Calibration is the first and best defense. Routine air calibration removes most rings; persistent rings signal drift or a failed element.⁹
• Software correction complements hardware. Polar-domain, total-variation, iterative, and learning-based methods suppress residual rings but should not mask a hardware failure.²⁵⁸
• Build ring detection into QC. Uniform-phantom review and uniformity/noise analysis catch developing rings before they reach clinical images.⁹

Conclusion

Ring artifacts are a direct, legible message from the detector: a channel at a known radius is not matching its neighbors. The geometry that turns a single bad channel into a concentric circle also makes the artifact easy to recognize and localize — the radius names the channel, and the completeness names the fault. The physicist's job is to catch rings in QC, judge their clinical and quantitative significance, and route them to the right fix: calibration for drift, service for failure, and software correction to recover affected studies. Treated this way, ring artifacts move from a recurring nuisance to a well-controlled, documented part of the CT quality program.¹²⁹

How DRPS Can Help

Diagnostic Radiation Physics Services helps imaging facilities keep CT image quality accurate and defensible. For ring artifacts and related scanner-based problems, this includes CT physics testing, acceptance and annual surveys, uniformity and CT-number QC, artifact investigation and localization, and accreditation support for ACR and state programs — all performed by board-certified medical physicists.

DRPS supports facilities across our service locations, including Florida, Maryland, Virginia, Washington DC, California, Nevada, New York, Pennsylvania, New Jersey, and Delaware.

A CT program that detects a ring, names the channel, and resolves the cause protects both image interpretation and the quantitative numbers clinicians rely on.

Related Resources

• CT image artifacts: causes and correction
• CT beam hardening artifacts and correction
• Metal artifact reduction in CT
• CT number (HU) calibration QC
• CT acceptance testing and commissioning
• CT physics testing services
• Accreditation support

References

1. Barrett JF, Keat N. Artifacts in CT: recognition and avoidance. RadioGraphics. 2004;24(6):1679-1691. doi:10.1148/rg.246045065. PubMed
2. Prell D, Kyriakou Y, Kalender WA. Comparison of ring artifact correction methods for flat-detector CT. Physics in Medicine and Biology. 2009;54(12):3881-3895. doi:10.1088/0031-9155/54/12/018. PubMed
3. Sijbers J, Postnov A. Reduction of ring artefacts in high resolution micro-CT reconstructions. Physics in Medicine and Biology. 2004;49(14):N247-N253. doi:10.1088/0031-9155/49/14/n06. PubMed
4. Eldib ME, Hegazy M, Mun YJ, Cho MH, Cho MH, Lee SY. A ring artifact correction method: validation by micro-CT imaging with flat-panel detectors and a 2D photon-counting detector. Sensors (Basel). 2017;17(2):269. doi:10.3390/s17020269. PubMed
5. Salehjahromi M, Wang Q, Zhang Y, et al. A new iterative algorithm for ring artifact reduction in CT using ring total variation. Medical Physics. 2019;46(11):4803-4815. doi:10.1002/mp.13762. PubMed
6. Šalplachta J, Zikmund T, Zemek M, et al. Complete ring artifacts reduction procedure for lab-based X-ray nano CT systems. Sensors (Basel). 2021;21(1):238. doi:10.3390/s21010238. PubMed
7. Zhao Y, Ma C, Ji DJ, et al. Structure-aware guided filtering for a ring artifact correction in synchrotron x-ray microtomography. Applied Optics. 2023;62(28):7400-7410. doi:10.1364/AO.501832. PubMed
8. Liu Y, Wei C, Xu Q. Detector shifting and deep learning based ring artifact correction method for low-dose CT. Medical Physics. 2023;50(7):4308-4324. doi:10.1002/mp.16225. PubMed
9. Samei E, Bakalyar D, Boedeker KL, et al. Performance evaluation of computed tomography systems: summary of AAPM Task Group 233. Medical Physics. 2019;46(11):e735-e756. doi:10.1002/mp.13763. aapm.org
10. Kak AC, Slaney M. Principles of Computerized Tomographic Imaging. New York: IEEE Press; 1988. slaney.org