Title :
Retrieval of Illegal and Objectionable Multimedia
Author :
Choi, ByeongCheol ; Kim, Jungnyeo ; Ryou, Jeacheol
Author_Institution :
Inf. Security Res. Div., ETRI, Seoul
Abstract :
Currently, requirements of the retrieval of illegal movies and objectionable multimedia (images and movies) are needed in Internet era. So we developed XMAS (X multimedia analysis system). XMAS consists of two service applications - One is a detecting solution objectionable multimedia for clean internet and another is a retrieval solution of illegal movies for digital rights. XMAS uses feature extraction method based on the MPEG-7 visual descriptors and machine learning and decision based on multi-class SVM (support vector machine). The experimental accuracy of XMAS is about 0.91~0.93 for images of objectionable multimedia and about 0.93 for retrieval of illegal movies.
Keywords :
Internet; feature extraction; information retrieval; learning (artificial intelligence); multimedia computing; support vector machines; video coding; Internet; MPEG-7 visual descriptors; X multimedia analysis system; digital rights; feature extraction; illegal movies; illegal multimedia retrieval; machine learning; multiclass SVM; objectionable multimedia retrieval; support vector machine; Codecs; Computer networks; Data mining; Feature extraction; Image retrieval; Information retrieval; MPEG 7 Standard; Motion pictures; Multimedia systems; Support vector machines; MPEG-7 Visual Descriptors; Multimedia; SVM;
Conference_Titel :
Networked Computing and Advanced Information Management, 2008. NCM '08. Fourth International Conference on
Conference_Location :
Gyeongju
Print_ISBN :
978-0-7695-3322-3
DOI :
10.1109/NCM.2008.140
