Junk image filtering method based on semi-supervision

Abstract

When image junk mails are detected and determined through junk image filtering technology research based on semi-supervision, text and images characteristics are extracted to conduct characteristic processing. Detecting and sorting are conducted by using obtained sorting models, new label sample are added continuously, a sorter is trained, sorting precision is improved, and simultaneously misjudgment is greatly reduced. A large amount of experimental data inspection shows that the method builds a high efficient junk mail webpage filtering system, high accuracy rate is guaranteed and simultaneously processing efficiency is greatly improved, and webpage detecting time is greatly shortened.

Description

technical field [0001] The present invention is a method of semi-supervised learning, using labeled picture samples, to train the support vector machine algorithm model, and to realize the detection of image-type spam, which mainly solves the detection efficiency of image-type spam in today's technology Problems such as low recall rate and low recall rate belong to the field of data mining and machine learning. Background technique [0002] The continuous improvement of text-based spam filtering technology drives spammers to explore new spam production techniques. As a result, image spam has become a popular spam communication medium today. According to McAfee's report in 2007, image spam accounts for about 30% of all spam. Image-type spam is to embed advertisements and other spam information into pictures in the form of text, and spread them wantonly to email clients as attachments or directly as the text content of emails. [0003] In 2007, Battista Biggio et al. propos...

AI Technical Summary

Problems solved by technology

Using edge features to detect image Spam can obtain an accuracy rate of 80%. The advantage of this type of classification algorithm is that using edge features can obtain text-intensive shape regularity and the amount of calculation is not large, but the disadvantage is that the text on the template Response to font changes is slow

Since it takes a lot of manpower and material resources to obtain labeled samples, it is relatively easier to obtain unlabeled samples.

Images