Image's MD5 on Twitter: how important is it?
- Thread starter Paskrtild
- Start date
-
- Tags
- md5 twitter twitter images
klique
Elite Member
- Mar 31, 2011
- 2,328
- 1,753
It can't hurt, and literally takes 2 minutes to change the md5 on 1000s of pictures.
- Jun 20, 2013
- 11,505
- 8,917
Check for hashmanger
Wilson Grant Fisk
Elite Member
- Nov 10, 2012
- 15,677
- 46,847
SpectroMX5
Junior Member
- Feb 8, 2008
- 158
- 210
Important. Changing MD5s are a no brainer, especially if you're doing automated posting of any kind for your clients. But that might not be enough.
Considering the urgency to protect growing ad revenues in today's current political climate, Twitter's been taking massive proactive steps in defending against spam and automation.
Read: https://blog.twitter.com/official/e...s-fighting-spam-and-malicious-automation.html
After all the shenanigans crypto scammers have been pulling by duping profile pics, Twitter's been forced to step up it's game specifically with image recognition. Matching meta data like an MD5 hash has very little computational expense, so that's an obvious first line of defense. But I wouldn't be surprised if they rolled out some of their improved image ID functions to other areas of the platform. Image recognition is computationally expensive, so would they run an advanced ID function on every image that gets posted every time? Doubt it. But they might start with lower trust/high risk accounts in a forward attempt to clamp down on farming.
Also relevant to note that Twitter acquired Madbits, an image recognition/search company back in 2014:
https://techcrunch.com/2014/07/29/twitter-acquires-image-search-startup-madbits/
The company is “primarily interested in the task of image search, and the creation of intelligent, dynamic image sets, to automatically organize large databases of images.”
So no one can say they don't have the resources or capability.
Considering the urgency to protect growing ad revenues in today's current political climate, Twitter's been taking massive proactive steps in defending against spam and automation.
Read: https://blog.twitter.com/official/e...s-fighting-spam-and-malicious-automation.html
After all the shenanigans crypto scammers have been pulling by duping profile pics, Twitter's been forced to step up it's game specifically with image recognition. Matching meta data like an MD5 hash has very little computational expense, so that's an obvious first line of defense. But I wouldn't be surprised if they rolled out some of their improved image ID functions to other areas of the platform. Image recognition is computationally expensive, so would they run an advanced ID function on every image that gets posted every time? Doubt it. But they might start with lower trust/high risk accounts in a forward attempt to clamp down on farming.
Also relevant to note that Twitter acquired Madbits, an image recognition/search company back in 2014:
https://techcrunch.com/2014/07/29/twitter-acquires-image-search-startup-madbits/
The company is “primarily interested in the task of image search, and the creation of intelligent, dynamic image sets, to automatically organize large databases of images.”
So no one can say they don't have the resources or capability.