【摘要】 MMS(Multimedia Messaging Service)作为一种新兴的移动数据业务,因其支持图片、声音、动画等多媒体的传输,越来越受到用户青睐。用户每天利用该服务定制,传送和转发大量的实时媒体信息。如何对这些信息进行过滤和自动分类,在现实中具有重要意义。通常过滤应用系统可以根据信息的源地址和目的地址实现该消息的正常路由或阻断。提出的方法则能够根据多媒体信息中的图片内容来决定是否路由或阻断,同时对这些图片按语义进行自动分类,为在MMS服务中过滤和分类未经请求或订阅的垃圾媒体信息(如商业广告等)提供了新的方案。更多还原

【Abstract】 With the fast development of multimedia message service (MMS), it has penetrated people’s life as an important way for realtime communication of message. Users use the service everyday to make, send and transmit huge amount of real-time multimedia messages. It’s important to filter and classify such information. Usually the filtering application can block messages by lists of addresses (black/white list) or use traditional text/keyword filtering methods. A new filtering scheme is proposed by using content-based image analysis technology, which can help filter trash messages in MMS,and at the same time,it can classify these image according to theirs semantic.更多还原