So much brain power only to be used to de-blur japanese p0rn
wooo!
Caught all the pedants :P
Recovering redacted information from pixelated videos
21–30 of 70 posts
Re: Recovering redacted information from pixelated videos
#22The article references the famous face-depixellating PULSE GAN. Let's remember that PULSE also gave us Barry O'Bama. https://www.google.com/amp/s/www.theverge.com/platform/amp/2...
You definitely shouldn't be downvoted, this is a very valid point.
Re: Recovering redacted information from pixelated videos
#23Yet again mosaic blur's inferiority to opaque rectangle is demonstrated.
I recently made use of my right to access personal information from someone who had sent me unsolicited marketing material.
The person who fulfilled my request sent me a PDF copy of their full customer list, where all entries had been blacked out except mine.
As you may anticipate, that blacking out was simply a black box drawn on top of the actual data. It took me all of 3 seconds to select all, copy, then paste in a word processor to confirm that the file contained personal data from hundreds of other individuals.
Re: Recovering redacted information from pixelated videos
#24Even the best can make obvious mistakes. "I hacked a hardware crypto wallet and recovered $2M [video]" https://news.ycombinator.com/item?id=30067340 Showing a blurry 16 out of 24 trezor wallet seed words https://youtu.be/dT9y-KQbqi4?t=1720
That probably isn't very useful since the video isn't moving.
Re: Recovering redacted information from pixelated videos
#25Earlier quoted context omitted.
I’m slightly curious as to how, in these cases, someone came to decide that blurring was the right way to do it. It seems unlikely that they never thought of simply blacking/ blanking the text, but if so, then blurring must have seemed preferable. My best guess is that, having seen blurring of faces (which is arguably OK when one merely wants to avoid casual attempts at identification, while retaining a ‘natural’ loo…
Retaining a "natural" look is phrased well. I think that's all behind the decision to infrasample (opposite of supersample?) parts of an image.
The opposite of supersample would be subsample.
Re: Recovering redacted information from pixelated videos
#26Yet again mosaic blur's inferiority to opaque rectangle is demonstrated.
Is anyone using this sort of idea on upscaling videos?
AI upscaling? Yes. Seems like some fans used AI upscaling for the older Star Gate SG-1 seasons, since these were direct-to-VHS and hence shot on low quality media (by now there are official bluray releases which are also somehow upscaled, but I was told those would be of inferior quality compared to the fan effort). Not sure if those efforts worked on a frame-by-frame basis or used information across frames.
[edit] here is a link: https://www.youtube.com/watch?v=cDss3BfsITs [/edit]
Re: Recovering redacted information from pixelated videos
#27The article references the famous face-depixellating PULSE GAN. Let's remember that PULSE also gave us Barry O'Bama. https://www.google.com/amp/s/www.theverge.com/platform/amp/2...
You definitely shouldn't be downvoted, this is a very valid point.
I assume I got down-voted by people cranky about AMP links.
Re: Recovering redacted information from pixelated videos
#28Earlier quoted context omitted.
You definitely shouldn't be downvoted, this is a very valid point.
I assume I got down-voted by people cranky about AMP links.
I basically only saw "Barry O'Bama" at first, and thought "oh good, another rightwingey troll/rant". I managed to pick up on the PULSE GAN before my mouse made it over to the downvote button, tho, so it wasn't me. I think you'd be better off just adding half a sentence of tl;dr like "... Barry O'Bama, where PULSE GAN incorrectly reconstructed a pixelated image of Barack Obama, resulting in a white guy that didn't look anything like the original" or something like that.
Re: Recovering redacted information from pixelated videos
#29Re: Recovering redacted information from pixelated videos
#30Even the best can make obvious mistakes. "I hacked a hardware crypto wallet and recovered $2M [video]" https://news.ycombinator.com/item?id=30067340 Showing a blurry 16 out of 24 trezor wallet seed words https://youtu.be/dT9y-KQbqi4?t=1720
That probably isn't very useful since the video isn't moving.
With a redaction as bad as a Gaussian blur at such a high resolution (and a dictionary of only 2048 candidates), you don't need multiple angles to decipher those words.
It'd be great if they replaced the key with bogus words first before blurring to troll people, but somehow I doubt it.