Detect image spam with content base information retrieval

at the present time, the increase of e-mail spam are heavy to cumber and the spam are vastly spread. These spams cause various problems to the Internet users, such as full incoming mailbox, and wasting time. Therefore, tremendous methods have been proposed but most of them have limitation in the mapping feature and processing time. This paper proposed a method that can detect a set of image e-mail spam. This proposes method can be described in 3 steps: firstly, this method receives an incoming mail and convert image to file and produce the characteristic of image database or corpus. Secondly, explain about the structure of the search corpus with Van Emde Boas trees structure and then retrieve data and verity image with content-bases image retrieval (CBIR) from corpus. The important feature of this research methodology is that it will sensitively distinguish e-mail message that has a partial similarity of e-mail spam from the normal e-mail. Therefore, Users will have a legitimate incoming mailbox as wish.