Talk:Image spam

Untitled
Doesn't the use of a genuine specimen of image spam on this page serve to propagate said spam?

Yeastbeast 03:17, 29 July 2007 (UTC)

Disputing claim
It is commonly claimed that the distortions are put in place to stop OCR.

I do not think this likely; the distortions are so minimal that OCR would still be easy.

I think a more likely explanation is that without distortions, email providers could quickly scan for all files matching an MD5 hash and mark them as spam. If the images have minor variations, their hashes will be unique and thus it makes it harder for providers to identify the spam emails.

I realise that wikipedia is not a place for original research, but thought it nonetheless noteworthy.

20 March 2013 Hi all, I have recently revised this page trying to address the open issues of the previous version, namely: 1) This article needs attention from an expert in Telecommunications. (November 2008) 2) This article does not cite any references or sources. (November 2009) I also tried to keep into account the major points raised above, that is, the problem of obfuscation and signature-based detection. Though I'm not an expert on Telecommunications, I did some major work on image spam detection. However, I could not understand why the old page is still visualized instead of the new one. Could anyone point out how the updating process takes place in this case, or provide any reference to understand how that works? Thanks! — Preceding unsigned comment added by 192.167.131.1 (talk) 13:55, 20 March 2013 (UTC)