dcmnorm
Read, inspect, edit, anonymize, convert, transcode, and render DICOM medical imaging files using the dcmnorm CLI, and send/receive/query them over a DICOM network using the dcmtalk CLI. Use this skill whenever the user mentions DICOM, .dcm files, medical images (CT/MR/US/DX/SR),
Install
npx skills add https://github.com/pohcee/dcmnorm/tree/main/plugin/skills/dcmnorm
claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install pohcee-dcmnorm@llmmart
git clone https://github.com/pohcee/dcmnorm.git
The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole pohcee/dcmnorm collection as a plugin from our marketplace. Git is the plain clone.
Skill manifest
dcmnorm — DICOM read / edit / write / render CLI
dcmnorm is a CLI that converts between DICOM and JSON, edits DICOM elements, transcodes transfer syntaxes, validates files, and renders pixel data to PNG/JPEG/raw/MPEG4. Make sure it's installed first (see Setup); run dcmnorm --help for the full option list.
Setup
This skill is written for dcmnorm 0.17.1. Before the first dcmnorm or dcmtalk command in a session, run dcmnorm --version (it prints e.g. dcmnorm 0.7.2-<build id>).
- Found, version 0.17.1 or newer: go ahead.
- Not found, or older: tell the user you're installing the dcmnorm command-line tools (native binaries from PyPI, nothing runs in Python), then install the pinned version with the first of these that's available:
If pip refuses withuv tool install --force dcmnorm==0.17.1 pipx install --force dcmnorm==0.17.1 python3 -m pip install dcmnorm==0.17.1externally-managed-environment: on the user's own machine, install uv or pipx instead; in a disposable sandbox, add--user --break-system-packages. This installs bothdcmnormanddcmtalk. Rundcmnorm --versionagain; if the command still isn't found, the installer's scripts directory isn't onPATH(uv/pipx print it; for pip it'spython3 -m site --user-baseplus/bin), so call the binary by its full path. - Install not possible (no network, no Python, a sandbox that blocks package downloads): say so plainly and point the user at the other install options at https://github.com/pohcee/dcmnorm#install (release archives,
.deb, Docker). Don't substitute another DICOM tool without asking.
Only the video transfer syntaxes (MPEG-2, MPEG-4 AVC/H.264, HEVC) and .mp4/.mov output also need the ffmpeg command. Check ffmpeg -version when a task needs one of those; if it's missing, dcmnorm fails with an install hint, so pass that on (apt install ffmpeg, brew install ffmpeg, winget install ffmpeg). dcmnorm --list-transfer-syntaxes shows what the installed build can decode and encode.
For talking to a DICOM network peer (PACS, modality, another node) instead of local files, use the companion dcmtalk CLI — see Network operations with dcmtalk below.
The core invocation is:
dcmnorm [OPTIONS] [INPUT] [OUTPUT]
The operation is inferred from the file extensions of INPUT and OUTPUT:
| INPUT | OUTPUT | Operation |
|---|---|---|
.dcm |
(none) | Print DICOM as JSON to stdout |
.dcm |
.json |
DICOM → JSON file |
.json |
.dcm |
JSON → DICOM file |
.dcm |
.dcm |
DICOM → DICOM (edit / transcode) |
.dcm |
.png / .jpg / .raw |
Render a frame to an image |
.dcm |
.mp4 / .mov |
Render frames to MPEG4 video |
.dcm (+ --mpr) |
.png / .dcm / .nii(.gz) / .nrrd |
Multiplanar reformation — see MPR below |
.dcm |
.gputex |
Lossless GPU texture export — see GPU texture export below |
When extensions are missing or misleading, override detection with --input-type dicom|json and --output-type dicom|json|raw|png|jpeg|mpeg4|texture.
Patient data
dcmnorm is a file and network toolkit, not a medical device: don't present its output as a basis for diagnosis or treatment. DICOM files usually carry protected health information (names, IDs, birth dates, accession numbers). Prefer synthetic or de-identified files, keep files on the user's machine, and don't repeat patient identifiers back in the conversation unless the user asks for them. De-identification with --set removes only what the edits name, so check the result (for example dcmnorm out.dcm | grep -i -E 'name|id|birth') before anyone shares it. dcmtalk sends data to whatever DICOM node the user names, so confirm the destination before C-STORE or C-MOVE.
Reading / inspecting DICOM
Print the whole dataset as flat JSON to stdout (bulk data like PixelData becomes a BulkDataURI reference, so output stays small):
dcmnorm file.dcm
Read only specific elements — much faster on big files since only filtered elements are parsed. Keys are DICOM keywords or GGGG,EEEE tags, repeated or comma-separated:
dcmnorm --filter PatientName,Modality,StudyInstanceUID file.dcm
# → {"Modality":"CT","PatientName":"Anonymized","StudyInstanceUID":"1.2..."}
JSON shape options (defaults: --format flat --keys name):
--format standard— PS3.18 DICOM JSON ({"00080060":{"vr":"CS","Value":["CT"]}})--keys hex— hex tags instead of keywords--bulk-data inline— base64-embed binary values instead of URI references (large output; avoid unless a full self-contained JSON is required)--omit-keyword-vm— with--format standard, leave out the non-standardKeywordandVMmembers (PS3.18 Annex F output)
Validating files
--check-dicom prints the path if the file is valid DICOM and exits non-zero otherwise:
dcmnorm --check-dicom file.dcm
Batch mode — pipe paths via -I / --stdin-paths; only valid DICOM paths are printed (great for finding DICOM files without relying on extensions):
find /data -type f | dcmnorm -I --check-dicom
-I also works for batch conversion/edit operations on many files.
--check-dicom-logic goes further: the file parses fine, but does its metadata actually make
sense (bits/samples/photometric relationships, pixel data length, transfer-syntax-vs-actual-bytes,
UID sanity, geometry)? Useful for catching a file that will fail or produce garbage output only
once you try to decompress/render/route it:
dcmnorm --check-dicom-logic --verbose file.dcm # OK/WARN/FAIL + per-finding detail
dcmnorm --check-dicom-logic --output-type json file.dcm # structured report
find /data -type f | dcmnorm -I --check-dicom-logic # batch, NDJSON with --output-type json
Errors (will plausibly break downstream decompress/render/routing) fail the exit code by default;
warnings (suspicious but usually survivable) only fail it with --strict.
Editing / updating DICOM
Set or remove elements (keywords or GGGG,EEEE tags, repeatable):
dcmnorm --set "PatientName=DOE^JOHN" --set "PatientID=12345" in.dcm out.dcm
dcmnorm --remove PatientBirthDate --remove "0010,1000" in.dcm out.dcm
dcmnorm --remove-private-tags in.dcm out.dcm # strip all private tags
Edit in place with --overwrite (no OUTPUT argument):
dcmnorm --overwrite --set PatientName=ANON file.dcm
Editing flags combine freely with each other and with conversion/transcoding in one pass. A typical de-identification:
dcmnorm --set PatientName=ANON --set PatientID=ANON \
--remove PatientBirthDate --remove-private-tags in.dcm out.dcm
Writing DICOM from JSON
JSON (flat or standard format, auto-detected) converts back to DICOM:
dcmnorm dataset.json out.dcm
For a lossless DICOM → JSON → DICOM round trip, export with a bulk-data source so binary values (PixelData etc.) can be resolved on the way back:
dcmnorm in.dcm dataset.json --bulk-data-source # embeds file:// URIs pointing at in.dcm
# ...edit dataset.json...
dcmnorm dataset.json out.dcm # URIs resolved automatically
Gotcha: --bulk-data-source takes an optional value. If a positional path follows it, the flag swallows that path as its value. When using the no-value form, put the flag after INPUT/OUTPUT (as above). With a value (--bulk-data-source /path/to/source.dcm, used on JSON→DICOM to resolve relative URIs), position doesn't matter.
Transcoding transfer syntaxes
dcmnorm --list-transfer-syntaxes # table of supported UIDs + decode/encode capability
dcmnorm --transfer-syntax 1.2.840.10008.1.2.1 in.dcm out.dcm
Common targets: 1.2.840.10008.1.2.1 Explicit VR Little Endian (uncompressed, maximally compatible), 1.2.840.10008.1.2.4.90 JPEG 2000 lossless, 1.2.840.10008.1.2.4.201 HTJ2K lossless (faster to encode/decode, ~3.5% larger), 1.2.840.10008.1.2.4.50 JPEG baseline. Check the PIXEL_ENCODE column in the list before targeting a compressed syntax. --jpeg2000-codec openjpeg|kakadu forces a specific classic JPEG 2000 decoder if needed; HTJ2K always decodes with OpenHTJ2K.
Visualizing / rendering
Render to PNG or JPEG (modality LUT, VOI windowing, and ICC profile applied automatically):
dcmnorm file.dcm out.png
dcmnorm --jpeg-quality 95 file.dcm out.jpg
Frame selection and video:
dcmnorm --render-frame 5 multi.dcm frame5.png # zero-based index
dcmnorm --render-all-frames multi.dcm frame.png # writes frame_000001.png, frame_000002.png, ...
dcmnorm --render-fps 30 cine.dcm loop.mp4 # fps defaults to DICOM metadata, else 24
Windowing and sizing:
dcmnorm --window-center 40 --window-width 400 ct.dcm soft-tissue.png
dcmnorm --scale-max-size 512 file.dcm thumb.png # longer side = 512, aspect preserved
dcmnorm --output-width 256 file.dcm out.png # height auto from aspect ratio
dcmnorm --pad --pad-color '#000000' file.dcm square.png # pad to square canvas
--no-voi-lut / --no-modality-lut / --no-icc-profile disable those pipeline stages when raw-er values are wanted.
Burn redaction boxes over the pixels (e.g. to mask burned-in PHI). Coordinates are in output pixels; negative X/Y anchor from right/bottom; W/H accept pixels or %; repeatable:
dcmnorm --redact-box 10,10,200,40 --redact-box -110,-60,100,50 \
--redact-color '#000000' us.dcm redacted.png
Multiplanar Reformation (MPR)
--mpr combines multiple slice files from one parallel stack (e.g. a CT/MR/PT series) into a 3D volume — honoring real ImagePositionPatient/ImageOrientationPatient/spacing, not just stacking images — and reformats a plane or a whole stack of planes out of it. Give it every slice file as INPUT with the last path as OUTPUT (shell globs work as usual):
dcmnorm --mpr axial series_dir/*.dcm axial.png
dcmnorm --mpr coronal series_dir/*.dcm coronal.png
dcmnorm --mpr 15,30,0 series_dir/*.dcm oblique.png # YAW,PITCH,ROLL degrees, oblique camera
Useful modifiers (all require --mpr): --mpr-origin X,Y,Z (mm, defaults to volume center), --mpr-depth MM (offset along the plane's own normal), --mpr-spacing MM (output pixel size), --mpr-thickness MM + --mpr-projection mip|minip|average (thick-slab projection instead of an infinitely-thin plane).
--mpr-depth also accepts a range (START:END, START:END:STEP, or all/all:STEP) to reformat a whole stack of slices instead of one plane — output extension decides the result shape: numbered PNGs, a proper multi-instance DICOM series (.dcm), or one whole-volume file (.nii/.nii.gz/.nrrd, float32 physical values, never 8-bit/windowed):
dcmnorm --mpr coronal --mpr-depth -40:40:2 series_dir/*.dcm coronal.dcm # multi-instance DICOM series
dcmnorm --mpr coronal --mpr-depth all --mpr-spacing 1 series_dir/*.dcm coronal.nii.gz
MPR mode is incompatible with --filter/--transfer-syntax/--set/--remove/--render-all-frames/--render-fps/--scale-max-size, and errors clearly (rather than silently mis-rendering) on an inconsistent/gantry-tilted stack.
GPU texture export (.gputex)
--output-type texture (or a .gputex output extension) packs a frame or volume as a lossless, GPU-upload-ready payload — the raw int16/uint16 sample lattice plus a rescaleSlope/rescaleIntercept pair for recovering physical values — instead of an 8-bit windowed render. Every export writes OUTPUT (payload bytes) plus OUTPUT.json (a metadata sidecar: dimensions, physical geometry, rescale slope/intercept, default window/level, compression):
dcmnorm frame.dcm frame.gputex # single frame, depth-1 texture
dcmnorm --mpr axial series_dir/*.dcm volume.gputex # whole series as one volume texture
dcmnorm --mpr axial series_dir/*.dcm volume.gputex --texture-compression none # skip gzip
--texture-max-dim <N> caps the longest axis (proportional downsample), for both a single-frame export and a --mpr volume export. A frame-stack variant (independent frames packed as a texture array, e.g. a cine loop's frames or one file per series instance, no resampling/physical geometry) exists only via each binding's own exportFrameStackTexture/export_frame_stack_texture (Node, Python, Java) — no CLI flag for it yet.
Tips
- After creating or editing a file, verify with
dcmnorm --check-dicom out.dcmand spot-check tags with--filter. - To view a DICOM image yourself (e.g. to answer "what does this scan show?" or verify a redaction), render it to PNG and read the image file.
- CT images often need explicit windowing to be readable: brain ≈ C40/W80, soft tissue ≈ C40/W400, lung ≈ C-600/W1500, bone ≈ C300/W1500.
--verboseprints conversion/rendering diagnostics when a command fails or produces unexpected output.- Detection handles files missing the standard preamble/meta group; for files with wrong or missing extensions, force it with
--input-type dicom.
Network operations with dcmtalk
dcmtalk is a CLI for DICOM network (DIMSE) operations — the same ground as dcmtk's echoscu/storescu/findscu/movescu/storescp. It's installed together with dcmnorm (see Setup); run dcmtalk --help or dcmtalk <subcommand> --help for the full option list. Every subcommand takes a peer address as HOST:PORT and accepts -v/--verbose to log association negotiation, presentation contexts, and each DIMSE command/response to stderr — reach for --verbose first whenever a connection to a real PACS behaves unexpectedly.
Check connectivity to a peer (C-ECHO):
dcmtalk echoscu pacs.example.com:11112
Send DICOM file(s) or a whole directory (recursive) to a peer (C-STORE):
dcmtalk storescu pacs.example.com:11112 study.dcm
dcmtalk storescu pacs.example.com:11112 ./study_dir/
Query a peer's studies (C-FIND) — keys are DICOM keywords, KEY=VALUE to match or bare KEY as a return key; matches print as one DICOM JSON line per study:
dcmtalk findscu pacs.example.com:11112 -k PatientID=12345 -k StudyDate
Ask a peer to push a study to another AE title it already knows (C-MOVE):
dcmtalk movescu pacs.example.com:11112 MY_STORE_AE 1.2.840.113619.2.55.3.604688119.971.1600000000.123
Retrieve one series (--level SERIES --series-uid <UID>) or specific instances (--level IMAGE --series-uid <UID> --sop-uid <UID> [--sop-uid <UID> ...]) of that study instead of the whole thing.
Run a temporary receiver to capture what a peer sends (a storescp), e.g. to inspect what a modality/PACS is actually transmitting — instances land under --cache-path as S_<StudyInstanceUID>/<Modality>_<SOPInstanceUID>.dcm, ready for dcmnorm to inspect:
dcmtalk storescp 11112 --ae-title MY_STORE_AE --cache-path ./received --verbose
dcmtalk storescp is receive-only: it answers C-FIND/C-MOVE requests with "unable to process" rather than serving a real query index or retrieve queue.
Most dcmtalk peers require the calling/called AE titles to match what they've been configured to accept — set them with -a/--calling-aet and -c/--called-aet if the default calling AE title (DCMTALK) or an unset called AE title gets an association rejected.
Files (dcmnorm)
-
SKILL.md 16 KB
--- name: dcmnorm description: Read, inspect, edit, anonymize, convert, transcode, and render DICOM medical imaging files using the dcmnorm CLI, and send/receive/query them over a DICOM network using the dcmtalk CLI. Use this skill whenever the user mentions DICOM, .dcm files, medical images (CT/MR/US/DX/SR), DICOM tags/metadata (PatientName, StudyInstanceUID, transfer syntax, SOP class), DICOM JSON, de-identifying/redacting medical images, wants to view/export DICOM pixel data as PNG/JPEG/MP4, or wants to talk to a PACS/DICOM node — connectivity checks (C-ECHO), sending files (C-STORE), querying (C-FIND), retrieving (C-MOVE), or receiving files (a storage SCP) — even if they don't name the dcmnorm or dcmtalk tools. Also use it to check whether files are valid DICOM or to batch-process directories of medical images. --- # dcmnorm — DICOM read / edit / write / render CLI `dcmnorm` is a CLI that converts between DICOM and JSON, edits DICOM elements, transcodes transfer syntaxes, validates files, and renders pixel data to PNG/JPEG/raw/MPEG4. Make sure it's installed first (see [Setup](#setup)); run `dcmnorm --help` for the full option list. ## Setup This skill is written for **dcmnorm 0.17.1**. Before the first `dcmnorm` or `dcmtalk` command in a session, run `dcmnorm --version` (it prints e.g. `dcmnorm 0.7.2-<build id>`). - **Found, version 0.17.1 or newer:** go ahead. - **Not found, or older:** tell the user you're installing the dcmnorm command-line tools (native binaries from PyPI, nothing runs in Python), then install the pinned version with the first of these that's available: ```bash uv tool install --force dcmnorm==0.17.1 pipx install --force dcmnorm==0.17.1 python3 -m pip install dcmnorm==0.17.1 ``` If pip refuses with `externally-managed-environment`: on the user's own machine, install uv or pipx instead; in a disposable sandbox, add `--user --break-system-packages`. This installs both `dcmnorm` and `dcmtalk`. Run `dcmnorm --version` again; if the command still isn't found, the installer's scripts directory isn't on `PATH` (uv/pipx print it; for pip it's `python3 -m site --user-base` plus `/bin`), so call the binary by its full path. - **Install not possible** (no network, no Python, a sandbox that blocks package downloads): say so plainly and point the user at the other install options at https://github.com/pohcee/dcmnorm#install (release archives, `.deb`, Docker). Don't substitute another DICOM tool without asking. Only the video transfer syntaxes (MPEG-2, MPEG-4 AVC/H.264, HEVC) and `.mp4`/`.mov` output also need the `ffmpeg` command. Check `ffmpeg -version` when a task needs one of those; if it's missing, dcmnorm fails with an install hint, so pass that on (`apt install ffmpeg`, `brew install ffmpeg`, `winget install ffmpeg`). `dcmnorm --list-transfer-syntaxes` shows what the installed build can decode and encode. For talking to a DICOM network peer (PACS, modality, another node) instead of local files, use the companion `dcmtalk` CLI — see [Network operations with `dcmtalk`](#network-operations-with-dcmtalk) below. The core invocation is: ``` dcmnorm [OPTIONS] [INPUT] [OUTPUT] ``` The operation is **inferred from the file extensions** of INPUT and OUTPUT: | INPUT | OUTPUT | Operation | |-------|--------|-----------| | `.dcm` | *(none)* | Print DICOM as JSON to stdout | | `.dcm` | `.json` | DICOM → JSON file | | `.json` | `.dcm` | JSON → DICOM file | | `.dcm` | `.dcm` | DICOM → DICOM (edit / transcode) | | `.dcm` | `.png` / `.jpg` / `.raw` | Render a frame to an image | | `.dcm` | `.mp4` / `.mov` | Render frames to MPEG4 video | | `.dcm` (+ `--mpr`) | `.png` / `.dcm` / `.nii(.gz)` / `.nrrd` | Multiplanar reformation — see [MPR](#multiplanar-reformation-mpr) below | | `.dcm` | `.gputex` | Lossless GPU texture export — see [GPU texture export](#gpu-texture-export-gputex) below | When extensions are missing or misleading, override detection with `--input-type dicom|json` and `--output-type dicom|json|raw|png|jpeg|mpeg4|texture`. ## Patient data dcmnorm is a file and network toolkit, not a medical device: don't present its output as a basis for diagnosis or treatment. DICOM files usually carry protected health information (names, IDs, birth dates, accession numbers). Prefer synthetic or de-identified files, keep files on the user's machine, and don't repeat patient identifiers back in the conversation unless the user asks for them. De-identification with `--set` removes only what the edits name, so check the result (for example `dcmnorm out.dcm | grep -i -E 'name|id|birth'`) before anyone shares it. `dcmtalk` sends data to whatever DICOM node the user names, so confirm the destination before C-STORE or C-MOVE. ## Reading / inspecting DICOM Print the whole dataset as flat JSON to stdout (bulk data like PixelData becomes a `BulkDataURI` reference, so output stays small): ```bash dcmnorm file.dcm ``` Read only specific elements — much faster on big files since only filtered elements are parsed. Keys are DICOM keywords or `GGGG,EEEE` tags, repeated or comma-separated: ```bash dcmnorm --filter PatientName,Modality,StudyInstanceUID file.dcm # → {"Modality":"CT","PatientName":"Anonymized","StudyInstanceUID":"1.2..."} ``` JSON shape options (defaults: `--format flat --keys name`): - `--format standard` — PS3.18 DICOM JSON (`{"00080060":{"vr":"CS","Value":["CT"]}}`) - `--keys hex` — hex tags instead of keywords - `--bulk-data inline` — base64-embed binary values instead of URI references (large output; avoid unless a full self-contained JSON is required) - `--omit-keyword-vm` — with `--format standard`, leave out the non-standard `Keyword` and `VM` members (PS3.18 Annex F output) ### Validating files `--check-dicom` prints the path if the file is valid DICOM and exits non-zero otherwise: ```bash dcmnorm --check-dicom file.dcm ``` Batch mode — pipe paths via `-I` / `--stdin-paths`; only valid DICOM paths are printed (great for finding DICOM files without relying on extensions): ```bash find /data -type f | dcmnorm -I --check-dicom ``` `-I` also works for batch conversion/edit operations on many files. `--check-dicom-logic` goes further: the file parses fine, but does its metadata actually make sense (bits/samples/photometric relationships, pixel data length, transfer-syntax-vs-actual-bytes, UID sanity, geometry)? Useful for catching a file that will fail or produce garbage output only once you try to decompress/render/route it: ```bash dcmnorm --check-dicom-logic --verbose file.dcm # OK/WARN/FAIL + per-finding detail dcmnorm --check-dicom-logic --output-type json file.dcm # structured report find /data -type f | dcmnorm -I --check-dicom-logic # batch, NDJSON with --output-type json ``` Errors (will plausibly break downstream decompress/render/routing) fail the exit code by default; warnings (suspicious but usually survivable) only fail it with `--strict`. ## Editing / updating DICOM Set or remove elements (keywords or `GGGG,EEEE` tags, repeatable): ```bash dcmnorm --set "PatientName=DOE^JOHN" --set "PatientID=12345" in.dcm out.dcm dcmnorm --remove PatientBirthDate --remove "0010,1000" in.dcm out.dcm dcmnorm --remove-private-tags in.dcm out.dcm # strip all private tags ``` Edit in place with `--overwrite` (no OUTPUT argument): ```bash dcmnorm --overwrite --set PatientName=ANON file.dcm ``` Editing flags combine freely with each other and with conversion/transcoding in one pass. A typical de-identification: ```bash dcmnorm --set PatientName=ANON --set PatientID=ANON \ --remove PatientBirthDate --remove-private-tags in.dcm out.dcm ``` ## Writing DICOM from JSON JSON (flat or standard format, auto-detected) converts back to DICOM: ```bash dcmnorm dataset.json out.dcm ``` For a lossless DICOM → JSON → DICOM round trip, export with a bulk-data source so binary values (PixelData etc.) can be resolved on the way back: ```bash dcmnorm in.dcm dataset.json --bulk-data-source # embeds file:// URIs pointing at in.dcm # ...edit dataset.json... dcmnorm dataset.json out.dcm # URIs resolved automatically ``` **Gotcha:** `--bulk-data-source` takes an *optional* value. If a positional path follows it, the flag swallows that path as its value. When using the no-value form, put the flag **after** INPUT/OUTPUT (as above). With a value (`--bulk-data-source /path/to/source.dcm`, used on JSON→DICOM to resolve relative URIs), position doesn't matter. ## Transcoding transfer syntaxes ```bash dcmnorm --list-transfer-syntaxes # table of supported UIDs + decode/encode capability dcmnorm --transfer-syntax 1.2.840.10008.1.2.1 in.dcm out.dcm ``` Common targets: `1.2.840.10008.1.2.1` Explicit VR Little Endian (uncompressed, maximally compatible), `1.2.840.10008.1.2.4.90` JPEG 2000 lossless, `1.2.840.10008.1.2.4.201` HTJ2K lossless (faster to encode/decode, ~3.5% larger), `1.2.840.10008.1.2.4.50` JPEG baseline. Check the PIXEL_ENCODE column in the list before targeting a compressed syntax. `--jpeg2000-codec openjpeg|kakadu` forces a specific *classic* JPEG 2000 decoder if needed; HTJ2K always decodes with OpenHTJ2K. ## Visualizing / rendering Render to PNG or JPEG (modality LUT, VOI windowing, and ICC profile applied automatically): ```bash dcmnorm file.dcm out.png dcmnorm --jpeg-quality 95 file.dcm out.jpg ``` Frame selection and video: ```bash dcmnorm --render-frame 5 multi.dcm frame5.png # zero-based index dcmnorm --render-all-frames multi.dcm frame.png # writes frame_000001.png, frame_000002.png, ... dcmnorm --render-fps 30 cine.dcm loop.mp4 # fps defaults to DICOM metadata, else 24 ``` Windowing and sizing: ```bash dcmnorm --window-center 40 --window-width 400 ct.dcm soft-tissue.png dcmnorm --scale-max-size 512 file.dcm thumb.png # longer side = 512, aspect preserved dcmnorm --output-width 256 file.dcm out.png # height auto from aspect ratio dcmnorm --pad --pad-color '#000000' file.dcm square.png # pad to square canvas ``` `--no-voi-lut` / `--no-modality-lut` / `--no-icc-profile` disable those pipeline stages when raw-er values are wanted. Burn redaction boxes over the pixels (e.g. to mask burned-in PHI). Coordinates are in output pixels; negative X/Y anchor from right/bottom; W/H accept pixels or `%`; repeatable: ```bash dcmnorm --redact-box 10,10,200,40 --redact-box -110,-60,100,50 \ --redact-color '#000000' us.dcm redacted.png ``` ## Multiplanar Reformation (MPR) `--mpr` combines multiple slice files from one parallel stack (e.g. a CT/MR/PT series) into a 3D volume — honoring real `ImagePositionPatient`/`ImageOrientationPatient`/spacing, not just stacking images — and reformats a plane or a whole stack of planes out of it. Give it every slice file as INPUT with the last path as OUTPUT (shell globs work as usual): ```bash dcmnorm --mpr axial series_dir/*.dcm axial.png dcmnorm --mpr coronal series_dir/*.dcm coronal.png dcmnorm --mpr 15,30,0 series_dir/*.dcm oblique.png # YAW,PITCH,ROLL degrees, oblique camera ``` Useful modifiers (all require `--mpr`): `--mpr-origin X,Y,Z` (mm, defaults to volume center), `--mpr-depth MM` (offset along the plane's own normal), `--mpr-spacing MM` (output pixel size), `--mpr-thickness MM` + `--mpr-projection mip|minip|average` (thick-slab projection instead of an infinitely-thin plane). `--mpr-depth` also accepts a range (`START:END`, `START:END:STEP`, or `all`/`all:STEP`) to reformat a whole stack of slices instead of one plane — output extension decides the result shape: numbered PNGs, a proper multi-instance DICOM series (`.dcm`), or one whole-volume file (`.nii`/`.nii.gz`/`.nrrd`, float32 physical values, never 8-bit/windowed): ```bash dcmnorm --mpr coronal --mpr-depth -40:40:2 series_dir/*.dcm coronal.dcm # multi-instance DICOM series dcmnorm --mpr coronal --mpr-depth all --mpr-spacing 1 series_dir/*.dcm coronal.nii.gz ``` MPR mode is incompatible with `--filter`/`--transfer-syntax`/`--set`/`--remove`/`--render-all-frames`/`--render-fps`/`--scale-max-size`, and errors clearly (rather than silently mis-rendering) on an inconsistent/gantry-tilted stack. ## GPU texture export (`.gputex`) `--output-type texture` (or a `.gputex` output extension) packs a frame or volume as a lossless, GPU-upload-ready payload — the raw `int16`/`uint16` sample lattice plus a `rescaleSlope`/`rescaleIntercept` pair for recovering physical values — instead of an 8-bit windowed render. Every export writes `OUTPUT` (payload bytes) plus `OUTPUT.json` (a metadata sidecar: dimensions, physical geometry, rescale slope/intercept, default window/level, compression): ```bash dcmnorm frame.dcm frame.gputex # single frame, depth-1 texture dcmnorm --mpr axial series_dir/*.dcm volume.gputex # whole series as one volume texture dcmnorm --mpr axial series_dir/*.dcm volume.gputex --texture-compression none # skip gzip ``` `--texture-max-dim <N>` caps the longest axis (proportional downsample), for both a single-frame export and a `--mpr` volume export. A frame-stack variant (independent frames packed as a texture array, e.g. a cine loop's frames or one file per series instance, no resampling/physical geometry) exists only via each binding's own `exportFrameStackTexture`/`export_frame_stack_texture` (Node, Python, Java) — no CLI flag for it yet. ## Tips - After creating or editing a file, verify with `dcmnorm --check-dicom out.dcm` and spot-check tags with `--filter`. - To view a DICOM image yourself (e.g. to answer "what does this scan show?" or verify a redaction), render it to PNG and read the image file. - CT images often need explicit windowing to be readable: brain ≈ C40/W80, soft tissue ≈ C40/W400, lung ≈ C-600/W1500, bone ≈ C300/W1500. - `--verbose` prints conversion/rendering diagnostics when a command fails or produces unexpected output. - Detection handles files missing the standard preamble/meta group; for files with wrong or missing extensions, force it with `--input-type dicom`. ## Network operations with `dcmtalk` `dcmtalk` is a CLI for DICOM network (DIMSE) operations — the same ground as dcmtk's `echoscu`/`storescu`/`findscu`/`movescu`/`storescp`. It's installed together with `dcmnorm` (see [Setup](#setup)); run `dcmtalk --help` or `dcmtalk <subcommand> --help` for the full option list. Every subcommand takes a peer address as `HOST:PORT` and accepts `-v`/`--verbose` to log association negotiation, presentation contexts, and each DIMSE command/response to stderr — reach for `--verbose` first whenever a connection to a real PACS behaves unexpectedly. Check connectivity to a peer (C-ECHO): ```bash dcmtalk echoscu pacs.example.com:11112 ``` Send DICOM file(s) or a whole directory (recursive) to a peer (C-STORE): ```bash dcmtalk storescu pacs.example.com:11112 study.dcm dcmtalk storescu pacs.example.com:11112 ./study_dir/ ``` Query a peer's studies (C-FIND) — keys are DICOM keywords, `KEY=VALUE` to match or bare `KEY` as a return key; matches print as one DICOM JSON line per study: ```bash dcmtalk findscu pacs.example.com:11112 -k PatientID=12345 -k StudyDate ``` Ask a peer to push a study to another AE title it already knows (C-MOVE): ```bash dcmtalk movescu pacs.example.com:11112 MY_STORE_AE 1.2.840.113619.2.55.3.604688119.971.1600000000.123 ``` Retrieve one series (`--level SERIES --series-uid <UID>`) or specific instances (`--level IMAGE --series-uid <UID> --sop-uid <UID> [--sop-uid <UID> ...]`) of that study instead of the whole thing. Run a temporary receiver to capture what a peer sends (a storescp), e.g. to inspect what a modality/PACS is actually transmitting — instances land under `--cache-path` as `S_<StudyInstanceUID>/<Modality>_<SOPInstanceUID>.dcm`, ready for `dcmnorm` to inspect: ```bash dcmtalk storescp 11112 --ae-title MY_STORE_AE --cache-path ./received --verbose ``` `dcmtalk storescp` is receive-only: it answers C-FIND/C-MOVE requests with "unable to process" rather than serving a real query index or retrieve queue. Most `dcmtalk` peers require the calling/called AE titles to match what they've been configured to accept — set them with `-a`/`--calling-aet` and `-c`/`--called-aet` if the default calling AE title (`DCMTALK`) or an unset called AE title gets an association rejected.
Comments (0)
Sign in to join the conversation.
Reviews (0)
No reviews yet.
No comments yet.