<title>DANCERS: Delft advanced news retrieval system</title>

Abstract

In this paper we present a system for automated analysis, classification and indexing of broadcast news programs. The system first analyzes the visual and the speech stream of an input news program in order to obtain an initial partitioning of the program into the so-called report segments. The analysis of the visual stream provides the boundaries of the report segments lying at the beginning and the end of each anchorperson shot. This analysis step is performed by applying an existing technique for anchorperson shot detection. The analysis of the speech stream gives the boundaries of the report segments lying in the middle of each (sufficiently long) silent interval. Then, the transcribed speech of each of the report segments is matched with the content of a large pre-specified textual topic database. This database covers a large number of topics and can be updated by the user at any time. For each topic a vast number of keywords is given, each of which is also assigned a weight that indicates the importance of a keyword for a certain topic. The result of the matching procedure is a list of probable topics per report segment, where for each topic on the list a likelihood is computed based on the number of relevant keywords found in the segment and on the weights of those keywords. The list of topics per segment is then shortened by separating the most probable from least probable topics based on their likelihood. Finally, the likelihood values of the most probable topics are used in the last system module to merge related neighboring segments into reports. The most probable topics serving as the base for the segment-merging procedure are at the same time the retrieval indexes for the reports and are used for classifying together all reports in the database that cover one and the same topic.

Video Summarization Using Mpeg-7 Motion Activity and Audio Descriptors

“…Topic detection has been mostly carried out using closed caption information, embedded captions and text obtained through speech recognition, by themselves or in combination with each other (See [8,9] for example). In such approaches, text is extracted from the video using some or all of the aforementioned sources and then processed using various heuristics to extract the topic(s).…”

ViBE: A Compressed Video Database Structured for Active Browsing and Search

“…Similarly, the pseudo-semantic distance is defined as the norm of the difference of the shot pseudo-semantic feature vectors as (9) We shall describe how these distance measures are used for browsing in Section V.…”

“…A number of systems have been proposed to solve the digital video library management problem. These include the VideoQ system [5], which allows region-based queries with motion information, the Virage video Engine [6], CueVideo from IBM [7] which uses keywords obtained from speech recognition to search through video data, a home movie library management system from Intel [8], Delft University's DANCERS news retrieval system [9] where image and audio data, and speech transcripts are used to divide news programs into report segments and assign topics to each of these and the Fischlar system [10] developed by the Dublin City University for indexing and browsing broadcast TV content. One of the largest efforts in video indexing and access is the Informedia system [11] developed by Carnegie Mellon University which reportedly has more than 2,000 hours of news and documentary programs.…”

Recent Advances in Video Content Analysis: From Visual Features to Semantic Video Segments

“…4 into a report. As an illustration, in the news segmentation approach proposed by Hanjalic et al, 35 first visual low-level features are used to detect the starting and ending points of all anchorperson shots. Several algorithms proposed in recent literature can be used for this purpose.…”