Levels of spam rose sharply in October to account for 89.07 per cent of all email during the month, according to figures from SoftScan.

The security firm warned that, unlike during the summer months which can create misleading statistics on spam levels as the amount of legitimate email drops, the latest rise is all due to additional spam.

Spam levels reached a record breaking 96 per cent in early October, supporting estimates that the trend over the whole month has been much higher than in the past two months.

"The spam highs normally seen in the summer are brought about by the low levels of normal business email as many people are away from work taking their summer holiday," said Diego d'Ambra, chief technology officer at SoftScan.

"But this time the levels of legitimate email have stayed the same and the spike is purely due to a marked increase of nearly four per cent in spam."

Virus levels remained low during October, however, accounting for just 0.41 per cent of email scanned by SoftScan. Phishing still remains the predominant threat.

Meanwhile, security vendor Secure Computing said that it has identified a 200 per cent increase in the amount of image spam over the past few months.

Image spam now accounts for 30 per cent of all spam and approximately one in every four messages circulating on the internet.

"Image-based spam is a particularly difficult problem for a couple of reasons," said Michael Osterman, founder and principal at research house Osterman Research.

"It is much harder to detect with conventional spam filtering and blocking technologies. It is also typically much larger than normal text-based spam, consuming much more bandwidth and storage."

Whereas traditional anti-spam software depends on content filtering techniques such as keyword filtering and Bayesian analysis to detect spam, these techniques are useless against image spam.

Even optical character recognition is not effective, according to Secure Computing.

However, Secure Computing claims that its TrustedSource platform is able to stop image spam as it collects information on email senders and the types of email they generate by accumulating data from more than 7,000 sensors in 48 countries.

As it gets more data, the breadth and depth of the TrustedSource database increases, accurately categorising sender reputations for multiple identities including IP, domain and URL reputations, according to the company.