Closed cleong110 closed 4 months ago
Recent, interesting paper.
pro: SOTA on How2Sign, self-supervised seems quite useful, potential code release. Also I recognize the authors and they've done interesting stuff. con: still a preprint, code not yet out.
https://github.com/cleong110/sign-language-processing.github.io/issues/11 my notes detailing things, incl. various drafts of the summary, iterating w/ help of ChatGPT, etc.
Has me thinking about if blurring really works for anonymization, isn't it reversible?
Recent, interesting paper.
pro: SOTA on How2Sign, self-supervised seems quite useful, potential code release. Also I recognize the authors and they've done interesting stuff. con: still a preprint, code not yet out.
My notes
https://github.com/cleong110/sign-language-processing.github.io/issues/11 my notes detailing things, incl. various drafts of the summary, iterating w/ help of ChatGPT, etc.
Has me thinking about if blurring really works for anonymization, isn't it reversible?