相容粗糙集在音频检索中的应用

Application of compatible rough sets in audio indexing

针对数字技术的发展促使个人的多媒体数据资料量越来越大,给人们的使用造成很大的不便,人们迫切需要能对这些资料进行检索和管理的工具,提出了利用相容粗糙集理论对音频数据进行预检索的新方法,即在相容类的近似空间中对音频数据进行预检索.实验结果表明,该方法可以有效地克服等价粗糙集在音频检索方面的限制,对提高检索效率具有一定的作用.

With the development of digital technology, the volume of personal multimedia archives are made ever-increasing, which brings much inconvenience to the users, so that it is urgent for the user to efficiently index and retrieve the multimedia. A novel method for audio data pre-indexing was presented with the help of compatible rough set, where the audio data were pre-indexed in a compatible-type approximate space. Experimental evaluation showed that the proposed approach could effectively overcome the limita- tion of equivalent rough set in respect of audio indexing and improve the efficiency of audio indexing to a certain extent.