Create a function that anonymize xrays

[elcronos]

Once we upload an image of an xray. We would need to create a function that removes the ID in that xray. We need the images uploaded to be anonymous if want to use it for future research work (Otherwise, we will not get Ethical Approval)

Resources:

• https://link.springer.com/article/10.1007%2Fs10916-017-0736-1
• https://hub.docker.com/r/bioinformaticsua/us-image-anonymizer

[timrosenow]

Easily done - I've already got a python class that does this. I'll modify and provide it.

[timrosenow]

I don't have permission to create a pull request or new branch. I have a file to upload, what is the best way to get it to you? Sorry i'm a bit of a git newbie :)

[Why-Ay-Es-Haitch]

I have utilized the EAST text-recognition model to put together a basic redaction tool to remove hardcoded text from image files. This needs to be done to mitigate privacy concerns. Unfortunately I don't have access to any x-rays with information on them to test the model. I have added the script at https://github.com/yash-parge/EAST-textRedaction for others to improve on.

[elcronos]

I just realized that AWS has textract. Which allows you to extract text from documents including images. The API also shows the position of the text in the image. Then, it will be easier to remove any text that could be in an image. I think this is the best solution. Anyone wants to work on this?

[JuanSeBestia]

Maybe remove the location in meta-info will be necessary?

Put random image for context

[timrosenow]

Yep I will From: Camilo Pestana <[email protected]> Sent: Monday, 23 March 2020 1:03 AM To: elcronos/COVID-19 <[email protected]> Cc: Tim Rosenow <[email protected]>; Comment <[email protected]> Subject: Re: [elcronos/COVID-19] Create a function that anonymize xrays (#19) I just realized that AWS has textract. Which allows you to extract text from documents including images. The API also shows the position of the text in the image. Then, it will be easier to remove any text that could be in an image. I think this is the best solution. Anyone wants to work on this? — You are receiving this because you commented. Reply to this email directly, view it on GitHub<#19 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ALJO6WDI76VWMQDGLFEUPKTRIZAFBANCNFSM4LQ4U6GA>.