EasyDICOM browser-viewing demonstration samples

The MR, CT and X-ray files are image-only demonstration reconstructions from the owner's
local DICOM corpus. The owner describes the supplied corpus as fully anonymized
public-domain images and explicitly authorized use for this local website demo.
This record documents that permission and the derivations; it is not a new
licence assertion for the complete original imaging corpus.

Sources (relative to the owner's data/dicoms/local directory):
  mr.dcm   : SERIES 2/000000.dcm       (MR brain image)
  ct.dcm   : SR/srdoc103/image13.dcm   (CT chest image)
  xray.dcm : Lateral_View0.dcm        (DX chest radiograph)

The OFFIS SR/srdoc103/readme.txt identifies image13.dcm as the GE CT #17 image
from RSNA '95 interoperability samples. The local MR and DX inputs have no
accompanying provenance or licence manifest in this corpus.

Processing:
  - New Explicit VR Little Endian Part 10 files, Secondary Capture SOP class.
  - Fresh deterministic demonstration study, series and instance identifiers.
  - Zero patient fields, dates, source UIDs, private tags, source descriptions,
    institutions, equipment identifiers, or original file-meta identifiers.
  - Only grayscale pixel layout, modality, numeric pixel spacing, default
    window, and rescale values are carried over from the source metadata.
  - MR and CT native pixel bytes are preserved exactly.
  - DX is nearest-neighbour reduced by 3 on each axis, using the centre sample
    of each 3-by-3 block (clipped at the final edge). Pixel spacing is multiplied
    by 3. Output: 886 columns by 1009 rows. Original stored sample values are
    retained for the chosen positions; this is a reduced demonstration image.
  - Images were visually reviewed before preparation. The selected images
    contain no readable identifying text. The DX retains an anatomical R
    marker. No automatic OCR or pixel redaction was applied.

All originals remain untouched and ignored in the original development kit.
Public website checks do not need that local corpus; they use these prepared
assets and independently generated synthetic fixtures.

Regenerate explicitly from the website directory (not part of the build):
  node scripts/prepare-viewer-samples.mjs /path/to/data/dicoms/local

The script asserts the grayscale native profile and dimensions, copies only an
explicit allowlist, and reparses each output to check its attributes and exact
pixel bytes. sample-manifest.json records technical dimensions, display values,
derivation, source references and SHA-256 hashes for the resulting files.

These are demonstration images, not a complete clinical DICOM IOD or viewer
validation set. The website viewer illustrates browser parsing and rendering.

Generated multi-frame sample:
  us.dcm is an independently generated ultrasound-style cine demonstration.
  It contains 24 distinct native 8-bit MONOCHROME2 frames, each 192 by 192 pixels.
  The speckled sector, moving structures and burned-in SYNTHETIC / US DEMO text
  are procedural artwork. No original imaging files or patient data are used.
  Frame Time is 50 milliseconds for a nominal 20 fps demonstration.
  The browser Play/Pause control decodes frames locally and loops at up to
  20 fps; playback of user files uses this demo speed, not acquisition timing.

  Rebuild it independently from the website directory:
    node scripts/generate-cine-sample.mjs

  Source: src/lib/synthetic-cine.js. The sample uses a generated Ultrasound
  Multi-frame Image Storage envelope and safe demonstration identifiers.
  It is illustrative artwork, not an acquired ultrasound image or diagnostic
  dataset. Existing MR/CT/X-ray provenance and sample-manifest.json are unchanged.