Segment organs and lesions in seconds, refine with full control, export anywhere — all in your browser
Manual contouring is one of the most time-consuming tasks in medical imaging: delineating an organ slice by slice can take longer than reading the entire study. DICOM Vision embeds state-of-the-art AI segmentation directly in its zero-footprint viewer, so you go from image to editable, exportable mask in seconds — with no software to install and no GPU workstation to maintain.
Segment interactively with SAM2 (Segment Anything Model 2): click on the structure you want, add positive or negative prompts to refine, and get a contour instantly. When you need the whole series done, domain-specific predictors for Lungs, Abdomen, Pancreas and Total Body segment every slice automatically. Results appear as color overlays with adjustable opacity, ready to inspect slice by slice.
An AI result you cannot edit is a dead end in clinical work. DICOM Vision pairs every segmentation with an interactive brush tool (1–100 px) to paint and erase masks, with full undo/redo. The dedicated editing workspace combines the 2D stack view with real-time 3D surface rendering, so every correction you make is immediately visible in three dimensions. The clinician always has the last word.
Masks are stored on the server with metadata — who created them, when, and with which properties — so you can reload and refine them in future sessions. When it is time to move on, export as DICOM RTSTRUCT for radiotherapy planning systems, NIfTI for research pipelines, or standard image formats for presentations and publications.
AI in DICOM Vision does not stop at masks. Automatic detection of anatomical landmarks — joint centers, vertebra centroids — feeds structured tables of clinical metrics and indices, with temporal tracking across longitudinal studies. Quantitative follow-up without a spreadsheet in sight.
All models run on DICOM Vision’s cloud infrastructure. Your side of the equation is a modern web browser: nothing to install, nothing to update, results in seconds on the same laptop or tablet you already use.
Interactive and automatic organ/lesion segmentation powered by SAM2 and specialized deep learning models.
Refine AI-generated segmentation masks with an interactive brush tool in a combined 2D+3D editing workspace.
Save segmentation masks to the server and export as DICOM RTSTRUCT, NIfTI, or image formats.
AI-powered detection of anatomical landmarks with automated extraction of clinical measurements and indices.
No. All AI models run on DICOM Vision's cloud infrastructure. You only need a modern web browser — segmentation results are computed server-side and streamed back to the viewer in seconds, on any laptop or tablet.
Dedicated predictors cover Lungs, Abdomen, Pancreas and Total Body segmentation across full series. For any other structure, SAM2 interactive segmentation lets you segment arbitrary organs and lesions with a few clicks.
Yes. An interactive brush tool (1–100 px) with undo/redo lets you paint and erase masks in a combined 2D+3D editing workspace, where corrections are instantly reflected in the 3D surface rendering.
Segmentation masks can be exported as DICOM RTSTRUCT for radiotherapy planning, NIfTI for research pipelines, or standard image formats. Masks are also stored on the server with metadata for later retrieval and refinement.
Yes. Masks are saved with the study, and DICOM Vision's share links let colleagues review the same series and overlays from their own browser, with no installation required.