dicom-pipeline
SkillDev toolsEnd-to-end DICOM workflow: parsing, anonymization/de-identification, conversion, structured reporting, PACS query/retrieve, and DICOMweb. Build automated medical imaging pipelines.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the dicom-pipeline skill
What this skill tells your AI
The instructions your AI receives, as published by mkurman/zorai in skills/scientific-skills/dicom-pipeline/SKILL.md and read by ahel’s review.
Overview
End-to-end DICOM workflow: parsing, anonymization, conversion, structured reporting, PACS query/retrieve, and DICOMweb integration.
Installation
uv pip install pydicom
Read and Inspect
import pydicom, numpy as np
ds = pydicom.dcmread("study.dcm")
print(f"Patient: {ds.PatientName}")
print(f"Modality: {ds.Modality}")
print(f"Study: {ds.StudyDescription}")
print(f"Size: {ds.Rows}x{ds.Columns}")
pixels = ds.pixel_array # NumPy array
Anonymization
ds = pydicom.dcmread("input.dcm")
phi_tags = [(0x0010, 0x0010), (0x0010, 0x0030), (0x0008, 0x0080)]
for tag in phi_tags:
if tag in ds:
ds[tag].value = ""
ds.save_as("anon.dcm")
DICOMweb
import requests
resp = requests.get(
"http://pacs:8080/dicom-web/studies",
params={"PatientName": "Doe*"},
headers={"Accept": "application/dicom+json"},
)
Workflow
- Parse DICOM with
pydicom.dcmread() - Extract metadata: modality, anatomy, patient info
- Anonymize per DICOM PS3.15 (clear PHI tags)
- Convert to NIfTI via dcm2niix or manual pixel_array
- Query PACS with DICOMweb QIDO-RS
- Generate DICOM SR (Structured Reports) for AI findings
Signals
- GitHub stars
- 324
- Forks
- 26
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
dicom-pipeline- Source
- github.com/mkurman/zorai