Image analysis software measures what is in a scan. It is not a viewer and it is not a PACS. A viewer shows the pixels. A PACS stores and retrieves them. Analysis software outlines a structure, returns a volume or a count, registers two visits, and writes a number a report can use.
If you meant the whole imaging stack — PACS, RIS, viewer, archive — that is medical imaging software. If you meant the algorithms (enhancement, segmentation, registration), that is medical image processing.
What the software actually does
Four jobs sit under the same label. Most products do one well and sell the rest as a screenshot.
| Job | What it is | What you get |
|---|---|---|
| Data | Open DICOM / NIfTI, keep the study together, talk to PACS | The right series, not a folder of loose JPEGs |
| Visualization | 2D cine, MPR, 3D surface / volume | A model you can rotate. Still not a measurement. |
| Algorithms | Segment, register, detect, classify | A mask, a transform, a box, or a label |
| Reporting | Numbers and a structured note back into the record | Volume, diameter, change vs last visit |
PYCAD Convert is a converter (DICOM ↔ NIfTI / PNG) with light 2D segmentation and 3D reconstruction. It is not analysis software. We build custom analysis viewers when the measurement has to live in a clinic app — see case studies.
Tools that are actually analysis software
The old list here sold FastStone (a photo browser), Optotherm (thermal cameras), Sunny (one vendor’s microscope), and Cytiva twice. Those are gone. These five are the ones people mean.
3D Slicer
Free, open-source, built for medical volumes. DICOM in, registration, segmentation, 3D view, a plugin tree that never ends. Research and pre-clinical work live here. The learning curve is real. Hardware wants RAM. Site: slicer.org.
ImageJ / Fiji
NIH public-domain Java. Fiji is the batteries-included build. Plugins for almost every 2D / light-3D lab job. The UI is 1990s. Huge stacks stall. A real walkthrough (threshold, watershed, Analyze Particles) is image segmentation in ImageJ, not this page. Site: imagej.net.
Image-Pro (Media Cybernetics)
Commercial 64-bit suite. Smart Segmentation (ML on faint objects), a data collector across a batch, life-science and materials labs that already paid for it. Price and a USB dongle are the tax. Site: mediacy.com/imagepro.
Icy
Open-source bioimage platform (Institut Pasteur). Visual workflows you wire as blocks; scripts in Java / Python when the block is not enough. Strong on cells and time-lapse. Heavy on big 3D. Site: icy.bioimageanalysis.org.
Digimizer
Windows measurement tool. Manual calipers plus simple object detection. Fast when the job is “length / area / count on a 2D image.” Not a DICOM workstation and not a 3D volume tool. Site: digimizer.com.
Also on a medical shortlist, not expanded here: ITK-SNAP (semi-auto 3D seg), Horos / OsiriX (macOS DICOM + some analysis), MITK (DKFZ toolkit). Wet-lab / blot / colony work is a different job — that is image lab software, not this list.
How they compare
| Tool | Cost | Medical DICOM / 3D | Use when |
|---|---|---|---|
| 3D Slicer | Free | Yes | Volumes, registration, research pipelines |
| ImageJ / Fiji | Free | Via plugins; 2D-first | Lab measurements, a scriptable 2D workflow |
| Image-Pro | Paid (dongle) | Some; life-science first | A lab that already standardized on it |
| Icy | Free | Bioimage, not a PACS client | Cells / time-lapse, visual workflows |
| Digimizer | Paid (Windows) | No | Quick 2D length / area / count |
PACS, EHR, RIS — the integration table
A tool that cannot talk to the archive does not get used. The 301’d “image analyzer” page was this table with more adjectives.
| Hook | What breaks | What “works” looks like |
|---|---|---|
| PACS | No query/retrieve; people export ZIP | DICOM C-FIND / C-MOVE (or a VNA) without a new worklist |
| EHR | Findings retyped into the chart | HL7 / FHIR / API writes a structured result to the record |
| RIS | Order and report never close | Bidirectional order / status so billing and the worklist match |
| Security | PHI on a laptop share | TLS in transit, encryption at rest, a BAA |
How to choose
- Modality and dimension. 2D micrographs are Fiji / Icy / Digimizer. CT / MRI volumes are Slicer (or a vendor workstation).
- Interoperability. Ask for a live query/retrieve against your PACS, not a slide. Analysis that cannot write back to the EHR is a side tool.
- Standalone vs suite. A specialty lab often wants one deep tool. A hospital wants one login. Depth vs one vendor is the real fork, not a feature matrix.
- Clearance. Clinical diagnosis in the US usually wants FDA-cleared software. Research can run Slicer. Do not mix those two sentences in a purchase order.
- Who clicks it. A tool the fellow will not open after week two is a sunk license.
FAQ
How is this different from a PACS?
PACS is the library: store, retrieve, show. Analysis software is the set of tools you use on a study once it is open — segment, measure, register, report. Some PACS vendors bolt analysis modules onto the viewer. That does not make the archive the analysis product. The whole stack (PACS + RIS + viewer) is covered on medical imaging software.
Do I need this if I already have a DICOM viewer?
A viewer window-levels and scrolls. Analysis software returns a mask or a number. If the report only needs “looks bigger,” a viewer is enough. If the report needs “12.4 ml, +8% vs June,” you need analysis.
PYCAD builds the imaging side of this when the measurement has to live in a clinic app. Case studies.