Medical imaging dataset · CT
Thin-slice, lossless non-contrast CT head.
More than two thousand head CT studies acquired on a consistent scanner platform, every one of them losslessly compressed and almost all reconstructed at 1.5 mm or finer. Nine in ten carry a radiologist report, and one in five patients is under eighteen, which is rare in openly available head CT.
- 2,100+
- Studies
- 800K+
- DICOM images
- 90%
- With radiologist report
- 100%
- Lossless compression
- 99%
- Thinnest series ≤ 1.5 mm
- 19%
- Paediatric (0–17)
Rounded inventory figures as of October 2026. Counts move as studies are indexed and qualified; exact eligible counts are returned against your specification.
Composition
Who is in the data
Demographics are carried as bands and recorded on every study profile, so a specification can include, exclude or balance by age and sex before anything is packaged.
Patient sex
- Male64%
- Female36%
Age band
- 0–1719%
- 18–3940%
- 40–5926%
- 60–7913%
- 80+2%
At a glance
- Series
- 7,800+
- Mean images per study
- ~375
- Acquisition years
- 2023 – 2026, over half in 2026
- Source institutions
- 2 high-volume centres
Percentages are of studies with the attribute recorded. Figures as of October 2026.
Acquisition profile
How the studies were acquired
Every attribute below is read from the DICOM headers and pixel data of each study, not from a catalogue description. These are the fields your specification can test against.
Contrast
94% non-contrast
A small contrast-enhanced subset, including arterial-phase studies, is flagged separately.
Slice thickness
≤ 1.5 mm in 99%
Sub-millimetre thinnest series in 22% of studies.
Tube voltage
130 kVp in 97%
Remainder at 110 kVp.
Bit depth
12 bits stored in 96%
Inside a 16-bit allocation, the standard export of the source fleet.
Scanner fleet
Siemens 100%
SOMATOM Scope 96%, SOMATOM go.Now 4%. A very consistent acquisition platform.
Coverage
~15 cm mean z-extent
Standard vertex-to-skull-base coverage; a small number of studies extend into the neck.
Qualification & delivery
Qualified against your specification, not sold by the terabyte
You do not buy the whole collection. You send inclusion criteria, we return the exact count of studies that pass, and only those are de-identified and delivered.
Specification
We turn your inclusion criteria into a machine-checkable requirement set: modality, anatomy, contrast phase, slice thickness, kVp, bit depth, compression, report pairing and demographic constraints.
Index & profile
Every study is read at header and pixel level and profiled on more than thirty attributes, including phase pattern, scanner model, coverage, compression history and report match.
Verdict per study
Each study receives an eligible, review or reject verdict with the exact criterion codes it failed. Studies whose phase cannot be read from metadata go to a radiologist, not a guess.
De-identify & deliver
Eligible studies are de-identified, re-verified, packaged with a hashed manifest and shipped to your bucket in batches. You load your own pass/fail list and we reconcile against it.
Status of this dataset
Head CT has not yet been delivered against a client specification, so the whole collection is available. Because the platform is so consistent, qualification is fast: tell us your slice, kVp, age and report requirements and we return the eligible count within days.
Read how the qualification service worksUse cases
What teams build with ct head data
Intracranial haemorrhage detection
Thin-slice, lossless non-contrast head CT at scale, with reports describing haemorrhage, infarct, mass effect and fracture where present.
Paediatric head CT
Around 400 studies in patients under eighteen, a population that is badly under-represented in public datasets.
Triage and worklist prioritisation
A mixed emergency and outpatient population for models that prioritise acute findings in a reading queue.
Report generation
Nearly two thousand paired reports for training and evaluating head CT report generation within the diagnostic scope.
Pixel-level labels, study-level classification and structured report extraction are available as a clinician-led annotation project. See medical data annotation.
Consent & permitted use
What this data may be used for, stated up front
Every study in the catalogue was obtained under a signed institutional research agreement with the source imaging network, with a data processing agreement on file. The permitted-use scope travels with the data and is written into every delivery manifest.
Permitted purpose
Development, validation and testing of diagnostic AI. Each delivery is matched to a declared purpose before a single study is packaged.
Permitted jurisdictions
United States and India. Deliveries to other jurisdictions require a fresh agreement with the source, which we will tell you plainly if it is not available.
Not permitted
General-purpose or foundation-model pretraining, and onward resale or sublicensing to third parties. We do not sell around the scope of the source agreement.
Retention
Time-limited retention is part of the agreement. Deliveries carry a retention term and we issue destruction certificates when a term ends or a study is withdrawn.
De-identification
DICOM headers are scrubbed to a documented profile, study, series and instance UIDs are remapped through a sealed crosswalk, and studies with burned-in annotation are detected and quarantined rather than shipped.
Provenance
Each batch ships with a hashed manifest, the acceptance verdict for every study and the source-agreement reference, so your counsel can trace the chain of rights.
Need the de-identification methodology and source-agreement summary for a vendor review? Request the compliance pack.
FAQ
CT Head dataset FAQ
Are findings labelled?
Not at pixel level by default. Each study ships with its radiologist report; study-level labels and slice-level or voxel-level annotation are available as a clinician-led annotation project scoped to your schema.
How is the paediatric subset handled?
Age is carried as a band, not a date of birth, and the paediatric subset is covered by the same institutional agreement. Tell us if your intended use requires excluding minors and we will qualify accordingly.
Does a two-centre source limit generalisation?
It limits scanner diversity, which we state plainly. It also means a very clean, consistent acquisition protocol. Many teams pair this collection with their own multi-centre data for domain-shift testing.
More datasets
Other collections in the catalogue
CTCT Abdomen & Pelvis
Multiphase CT abdomen & pelvis, qualified study by study.
- 4,300+
- Studies
- 5.8M+
- DICOM images
- 93%
- With radiologist report
CTCT Chest
Thin-slice CT chest across an older adult population.
- 1,250+
- Studies
- 950K+
- DICOM images
- 83%
- With radiologist report
MRIMRI Spine & MSK
Spine and joint MRI from a routine outpatient population.
- 600+
- Studies
- 115K+
- DICOM images
- 72%
- With radiologist report
Send us your inclusion criteria.
We will return the exact number of studies in this collection that pass, what the rest fail on, and a sample batch your pipeline can load, before any commercial discussion.
Or email [email protected] · Mohali, India