Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Image Scrubber: tool for anonymizing photographs taken at protests

everestpipkin.github.io

Post reply on HN
Print view View original

401–410 of 410 posts

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#401

Post by tehjoker »

jtbayly wrote: post #188
tehjoker wrote: post #109

Earlier quoted context omitted.

It's not really. Police lie constantly, especially when it would implicate them. Their statements carry zero information except where they provide positive publicly verifiable evidence. Unsettling and sad, but its the truth.

So I’ll take that as a “No, there’s no evidence whatsoever that these are connected. It’s entirely baseless speculation and conspiracy theory.”

Given the circumstances, I would say you're never going to know the truth about these deaths with the possible exception of documents declassified or admissions of guilt decades after the fact. The question is if you are a political activist, is do you take these deaths seriously even if you can't prove they are foul play? Given the record of the government, I just can't rule it out.

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#402

Post by tgsovlerkhgsel »

Hello71 wrote: post #302
tgsovlerkhgsel wrote: post #291

Earlier quoted context omitted.

Because the result is a lot more ugly. Don't tell people what not to do. Figure out why they're doing it, and provide what they actually want while still achieving the goals (here: security). Very coarse mosaic, add noise, then blur seems reasonably safe, and doesn't have to look like crap.

"seems reasonably safe" seems like a terrible cryptographic analysis. in fact, given that we already know that both blurring and mosaicing are individually reversible, and noise is easily removable from a sufficiently wide mosaic, this seems like a particularly terrible algorithm. that's not the point though: any man can create an encryption algorithm that he himself cannot break. maybe you can come up with an obfusc…

I was writing a HN comment, not a scientific paper, which is why I wrote "seems safe" instead of making stronger claims.

I'd also like to know how mosaicing is reversible, since it demonstrably reduces the total available amount of information from e.g. 20x20 = 400 RGB values to a single RGB value. This is not sufficient for text where you can start brute-forcing individual options because the search space is small and inputs can be reconstructed precisely, but I'd like to see an explanation why you think this is reversible for photos (even without noise added). I'd also like to know how you want to remove random noise applied to each mosaic block.

The mosaicing is supposed to be the security step here. The blur is optional eye candy not expected to remove further information.

In particular, if you claim that a face mosaiced with a large "pixel" size (e.g. so that the typical face is 5x5 "mosaic blocks" big), you're effectively claiming that you can perform facial recognition based on noisy 5x5 pixel images.

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#403

Post by heavyset_go »

abathur wrote: post #313
heavyset_go wrote: post #307

Earlier quoted context omitted.

A bad actor could selectively serve a different version to those they want to target.

There's no way to obtain and execute source code that you didn't write and hand-compile for which this risk doesn't exist. (And it applies in its own sense to books, paintings, phone calls from mom, letters from an old mentor, DVDs, rental cars, ...)

Download, verify keys and signatures. You could run a checksum or even read the code yourself depending on how paranoid you are. Otherwise, you're just hoping mycrimepics.net/dontsnitch wasn't subpoenaed between your last visit and now.

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#404

Post by pgalvin »

mceachen wrote: post #246
pgalvin wrote: post #121

Earlier quoted context omitted.

> metadata gets scrambled by both their "high quality" and "original" settings. This is NOT true, at least for the "original" setting. Upload an image, download it again - the checksums are identical. If you modify the time/date or add comments within Google Photos, that new information is kept in a .json file instead of the exif data, but Google NEVER modifies the original photo if you select "original" quality.

That's not my experience with a couple Google Pixels, using original quality. If you fetch your image via takeout or via the API, the image bytes are the same, but the EXIF headers have been changed. I've tried in both a standard Gmail account and a custom domain. Same behavior. Are you saying your file SHA is the same from your device and your takeout? I wonder if I am seeing US-account-exclusive behavior?

Yep, the SHA1/256 checksum of the original on my Android device is the same as the one from both Google Takeout and downloading from the Google Photos Web UI.

Same for the .heic files on my iPhone, too.

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#405

Post by mceachen »

pgalvin wrote: post #404
mceachen wrote: post #246

Earlier quoted context omitted.

That's not my experience with a couple Google Pixels, using original quality. If you fetch your image via takeout or via the API, the image bytes are the same, but the EXIF headers have been changed. I've tried in both a standard Gmail account and a custom domain. Same behavior. Are you saying your file SHA is the same from your device and your takeout? I wonder if I am seeing US-account-exclusive behavior?

Yep, the SHA1/256 checksum of the original on my Android device is the same as the one from both Google Takeout and downloading from the Google Photos Web UI. Same for the .heic files on my iPhone, too.

Interesting, thanks for sharing. I wonder what's different between our accounts and devices.

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#406

Post by tornato7 »

_bxg1 wrote: post #233
Ansil849 wrote: post #119

Some tips to maximise user privacy while deploying this tool: 1) The code, for now, runs locally. This is good. To avoid the possibility of the code being tampered with at a later day (for example, it could be modified to send copies of the image to a server), download the webpage and use the saved copy, not the live copy. 2) Do not use the blur functionality. For maximum privacy, this should be removed from the app…

A replacement for blur could just be black boxes. Seems easy and safe enough.

Even better, replace the face with Kim-Jong Un and then blur it. When they apply those forensic techniques they'll discover that it was him all along!

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#407

Post by Hello71 »

tgsovlerkhgsel wrote: post #402
Hello71 wrote: post #302

Earlier quoted context omitted.

"seems reasonably safe" seems like a terrible cryptographic analysis. in fact, given that we already know that both blurring and mosaicing are individually reversible, and noise is easily removable from a sufficiently wide mosaic, this seems like a particularly terrible algorithm. that's not the point though: any man can create an encryption algorithm that he himself cannot break. maybe you can come up with an obfusc…

I was writing a HN comment, not a scientific paper, which is why I wrote "seems safe" instead of making stronger claims. I'd also like to know how mosaicing is reversible, since it demonstrably reduces the total available amount of information from e.g. 20x20 = 400 RGB values to a single RGB value. This is not sufficient for text where you can start brute-forcing individual options because the search space is small a…

according to https://www.bitsoffreedom.nl/2019/12/12/amazons-rekognition-..., current facial recognition software is able to distinguish faces in very blurry mosaics using statistics.

it doesn't matter though. as I've explained, it's far easier to come up with flawed schemes than prove them insecure. just because I can't explain why your specific scheme is insecure doesn't mean it stands a chance against real cryptographers.

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#408

Post by tgsovlerkhgsel »

Hello71 wrote: post #407
tgsovlerkhgsel wrote: post #402

Earlier quoted context omitted.

I was writing a HN comment, not a scientific paper, which is why I wrote "seems safe" instead of making stronger claims. I'd also like to know how mosaicing is reversible, since it demonstrably reduces the total available amount of information from e.g. 20x20 = 400 RGB values to a single RGB value. This is not sufficient for text where you can start brute-forcing individual options because the search space is small a…

according to https://www.bitsoffreedom.nl/2019/12/12/amazons-rekognition-... , current facial recognition software is able to distinguish faces in very blurry mosaics using statistics. it doesn't matter though. as I've explained, it's far easier to come up with flawed schemes than prove them insecure. just because I can't explain why your specific scheme is insecure doesn't mean it stands a chance against real crypto…

The 20x26 example is indeed scary, but in line with what was known about facial recognition. (It also becomes a bit less scary when you don't look at a zoomed-in version of the image.)

Hence my suggestion to reduce a face to something like 5x5 blocks.

While I'm familiar with the crypto design problem, this is not a crypto algorithm. Sure, it can't be ruled out that someone in the future will find a way to do it, but the state of the art says that 5x5 pixels are not anywhere near enough to run face recognition.

And a solution that may be broken in the future is often much better than a solution that people don't use because it doesn't meet their needs, which in this case is not having fugly black boxes in their picture.

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#409

Post by fwip »

sfj wrote: post #278
fwip wrote: post #266

Earlier quoted context omitted.

Your first link shows: more black people are killed per-capita by cops than white people. Y Your second link shows: Black people are more likely to be convicted of homicide than white people. You've made a number of unfounded leaps on top of that: That "Black people commit 52% of crime." This is wrong in several ways. You're assuming that conviction rates accurately represent rates of crimes, and also that "all crime…

> Your first link shows: more black people are killed per-capita by cops than white people. Y Your second link shows: Black people are more likely to be convicted of homicide than white people. Yes, that's right. See the connection between crime and cops killing civilians? > But instead, we see that black people are convicted of only 27% of overall crime. Where do you see this? I can't find this anywhere. > That's in…

> Where do you see this? I can't find this anywhere.

Maybe check real sites instead of your usual racist ones? Google is a good start. Here's the first link! https://www.sentencingproject.org/publications/un-report-on-...

Re: Image Scrubber: tool for anonymizing photographs taken at protests

#410

Post by Hello71 »

tgsovlerkhgsel wrote: post #408
Hello71 wrote: post #407

Earlier quoted context omitted.

according to https://www.bitsoffreedom.nl/2019/12/12/amazons-rekognition-... , current facial recognition software is able to distinguish faces in very blurry mosaics using statistics. it doesn't matter though. as I've explained, it's far easier to come up with flawed schemes than prove them insecure. just because I can't explain why your specific scheme is insecure doesn't mean it stands a chance against real crypto…

The 20x26 example is indeed scary, but in line with what was known about facial recognition. (It also becomes a bit less scary when you don't look at a zoomed-in version of the image.) Hence my suggestion to reduce a face to something like 5x5 blocks. While I'm familiar with the crypto design problem, this is not a crypto algorithm. Sure, it can't be ruled out that someone in the future will find a way to do it, but…

this seems like a false dichotomy. there's nothing stopping you from using a pink oval, as long as it covers the face.
Post reply on HN