TY  - THES
T1  - iQuery a search engine for CNN news broadcasts using multimedia indexing and speech recognition
A1  - Pablo, Pauline Ann N.
A2  - Santos, Ariane Acena L.
A2  - Sebido, Joven J.
LA  - English
UL  - https://tuklas.up.edu.ph/Record/UP-99796217609396368
AB  - iQuery is a search engine for CNN news broadcasts which aims to redefine the method by which users search through news videos on the Internet. Its goal is to increase the relevance of search results returned to users whenever they query the news database for news clips using textual search parameters.     Unlike other search engines, the iQuery news darabase is populated by an intelligent Mulitmedia Indexing (MMI) Module [6][12][13][20]. Given a continuous video containing two or more news stories, this module detects when a news anchor appears in the video. After this, the module automatically segments the video into news clips following the news anchor - news story - news anchor format [3][19]. Keywords are then assigned to each segmented news clip. These keywords are based on the text transcribed by the speech recognizer [11], but with stop words omitted. These segmented news clips, along with their keywords, are stored in an online database.     iQuery allows administrators of the system to populate the database with the results of the MMI Module by means of a desktop application installed on a client computer. This desktop application processes the continuous news video, uploads the segmented news clips to an online server via File Transfer Protocol (FTP) and updates the online database with the ID and the keywords of each segmented news clip. Regular users, on the other hand, are presented with a web interface wherein textual search parameters are entered and relevant news clips, i.e. news clips containing the search parameter in their list of keywords, are retrieved.
CN  - LG 993.5 2010 C65 P33
KW  - Multimedia (Art) --Abstracting and indexing.
KW  - Automatic speech recognition.
KW  - Search engines --Computer programs.
ER  -