Introduction to image quality – Learn Radiology Physics

In diagnostic radiology, the ultimate goal is to produce an image that enables accurate diagnosis at the lowest reasonably achievable radiation dose.

Image quality therefore refers not simply to how “clear” an image appears, but to how effectively it displays clinically relevant information.

What Is Image Quality?

Image quality is the degree to which an image faithfully represents the patient’s anatomy and pathology.

It is influenced by both physical factors (detector performance, beam quality, exposure) and perceptual factors (how the eye and brain interpret the image).

A high-quality image is one that:

Demonstrates all relevant anatomical detail.
Has adequate contrast to differentiate tissues of interest.
Contains minimal noise and artefacts.
Was obtained with an appropriately low patient dose.

In practice, image quality is a trade-off between diagnostic adequacy and radiation exposure.

Determinants of Image Quality

The appearance and diagnostic value of an image are governed by several measurable parameters. we will look at these in more depth throughout this module.

Parameter	Definition	Determined By
Spatial resolution	Ability to distinguish small structures that are close together.	Detector pixel size, sampling frequency, focal spot size, motion.
Contrast resolution	Ability to display small differences in X-ray attenuation.	Bit depth, SNR, kVp, scatter.
Noise / SNR	Random variation in signal that obscures detail.	Photon statistics (quantum mottle), electronic noise, dose.
Sharpness / unsharpness	Clarity of edges and fine detail.	Focal spot geometry, movement, detector blur.
Artefacts	Features not present in the object imaged.	Equipment faults, processing errors, operator factors.

These parameters are inter-related: improving one often compromises another (for example, reducing noise by increasing dose).

The Image Quality–Dose Relationship

The relationship between image quality and patient dose is non-linear:

Increasing dose improves SNR and CNR, but with diminishing returns.
Above a certain threshold, further dose increases yield little improvement in perceived quality.
Below that threshold, noise dominates and diagnostic confidence falls sharply.

This relationship underpins the optimisation principle: achieving images that are “as good as necessary, not as good as possible”.

SNR ∝ √Dose

Doubling the dose improves SNR by only about 41 %. Therefore, dose increases should be carefully justified by diagnostic need.

The Imaging Chain

Every step from X-ray generation to display contributes to final image quality. Degradation at any step cannot be fully corrected later.

X-ray production: spectrum, filtration, and beam quality.
Patient interaction: attenuation, scatter, and subject contrast.
Detection: DQE, pixel size, sampling, noise.
Processing: corrections, windowing, enhancement.
Display: monitor resolution, viewing conditions.

Objective vs Subjective Image Quality

Objective measures:
Quantifiable metrics such as MTF, DQE, SNR, and NPS. Used in system testing and quality assurance. More on these later.
Subjective measures:
Human perception, image appearance, diagnostic acceptability, observer confidence.

Both are essential.

A technically perfect image may not improve diagnosis if it doesn’t enhance relevant anatomical visibility.

Key Takeaways and Exam Tips:

Image quality reflects how well an image serves its diagnostic purpose at minimal dose.
Key determinants: spatial resolution, contrast resolution, SNR, sharpness, and artefacts.
Always interpret quality in relation to clinical task.
Improving image quality usually increases dose. Optimisation seeks balance.
Common exam question: “List the main factors affecting image quality.”