PixelGuard is an advanced software solution designed to deidentify medical images while preserving clinical relevance. It ensures patient confidentiality by removing identifiable data embedded within pixel-level data and metadata. Its AI-enabled technology supports multiple languages for removing pixel data.
PixelGuard accepts a wide range of medical imaging formats, including:
Images can be saved in a number of output formats:
Note that a JPEG (.jpg) or PNG (.png) will not be transformed from a lower information state to a higher one (i.e. DICOM .dcm). PixelGuard generates deidentified medical images in the same format as the input. Key features include:
Yes. PixelGuard adheres to requirements including HIPAA, GDPR, and other applicable standards. Its processes are customizable to align with the needs of the end user.
Yes. PixelGuard is optimized for scalability and can process datasets containing thousands or even millions of images efficiently by leveraging parallel processing via elastic scaling and batch workflows. It is also set to handle jobs asynchronously.
Yes. Users can review cases (e.g., detected burned-in PHI) to ensure that the appropriate information has been redacted.
While PixelGuard is primarily cloud-based, we can work directly with the end user to ensure flexibility based on organizational needs.
PixelGuard employs advanced algorithms to maintain the diagnostic quality of images while securely redacting PHI. This includes decompression and re-compression as appropriate, safeguarding against lossy compression.
Yes. PixelGuard can select data from a user's cloud account as well as upload from specific folder/file locations on a local machine, enabling seamless integration with Picture Archiving and Communication Systems (PACS) or Radiology Information Systems (RIS).
Additionally, if users have AWS cloud solution architects available as part of their enterprise solutions, they can assist in getting you started.
PixelGuard is available on the AWS Marketplace. You may also contact our team through our website or email to request a demo, pricing information, or installation instructions.
Yes. Users can customize deidentification settings to fit their specific compliance workflows and organizational needs. This includes selecting which metadata to redact, swap, shift, etc. We also offer various algorithm versions that not only deidentify but also prepare data for AI-based training, with options for shifting, swapping, random assignment, or user-uploaded values for specific PatientIDs or dates.
Yes! Our solution records any redaction, swapping, or shifting of Patient IDs, dates, etc. in a crosswalk file that includes original filenames and new filenames (if applicable based on user configuration) along with the redacted text.