Image Anonymization Module

Overview

The imageAnonymization module provides functionalities for anonymizing DICOM images by replacing metadata values with randomly generated or de-identified values. This ensures that sensitive patient information is removed while preserving the structural integrity of the image data.

API Reference

anonymize

Anonymizes a given DICOM series by replacing all metadata with randomized values. The function iterates through all image instances in the series and modifies their metadata fields.

Syntax

anonymize(series: Series): Series

Parameters

Parameter	Type	Description
series	Series	The DICOM series to be anonymized.

Returns

Series - The anonymized DICOM series.

Example

import { anonymize } from "./imageAnonymization";
const anonymizedSeries = anonymize(originalSeries);

Implementation Details

• Iterates through all image instances in the series.

• Replaces metadata values based on their value representation (VR) type.

• Updates key metadata attributes such as:

  Parameter	Description
  seriesUID	Index of the current image in the series
  instanceUID	Index of the current image in the series
  studyUID	Index of the current image in the series
  accessionNumber	Index of the current image in the series
  studyDescription	Index of the current image in the series
  patientName	Index of the current image in the series
  patientBirthdate	Index of the current image in the series
  seriesDescription	Index of the current image in the series

• Marks the series as anonymized (series.anonymized = true).

Internal Utility Functions

• makeDeIdentifiedValue

makeDeIdentifiedValue(length: number, vr: string): string | undefined

Generates a de-identified value based on the provided length and value representation (VR):

1. LO, SH, PN → Replaced with a random alphanumeric string.

2. DA (Date) → Set to 19000101 (January 1, 1900).

3. TM (Time) → Set to the current time with zero-padding.

• makeRandomString

makeRandomString(length: number): string

Generates a random alphanumeric string of a given length.

• pad

  pad(num: number, size: number): string

Pads a number with leading zeros to ensure a fixed-size string representation.

Warnings & Considerations

This module does not alter pixel data, only metadata. Always ensure compliance with relevant data privacy regulations (e.g., HIPAA, GDPR) before anonymizing and sharing DICOM images. This module is provided under an open-source license. Use it responsibly in compliance with local regulations and privacy policies.