Multimodal Indexing of Multilingual News Video

Abstract

The problems associated with automatic analysis of news telecasts are more severe in a country like India, where there are many national and regional language channels, besides English. In this paper, we present a framework for multimodal analysis of multilingual news telecasts, which can be augmented with tools and techniques for specific news analytics tasks. Further, we focus on a set of techniques for automatic indexing of the news stories based on keywords spotted in speech as well as on the visuals of contemporary and domain interest. English keywords are derived from RSS feed and converted to Indian language equivalents for detection in speech and on ticker texts. Restricting the keyword list to a manageable number results in drastic improvement in indexing performance. We present illustrative examples and detailed experimental results to substantiate our claim.

A Unified System for Processing News and Detecting Advertisements in TV Streams

“…The topics segmentation for news program has been studied by many researchers such as in [34], [12], [32], [9], [28]. It has processed in two techniques: based on audio transcription methods and based on text detection methods.…”

“…Manson et al [19] proposed a metadata-based system that automatically structures TV stream, the system first detects inter-programs as repeated sequences in the broadcasted stream, and then deduce the boundaries of programs. In audiovisual-based systems, some solutions have been proposed that use speech transcription to extract information like proposed by [12], [24], and [21].…”

Shot Boundary Detection for Gujarati News Video

“…Another problem is transcriptions for Indian videos are not available. Due to all such aspects, text-based approach is not found efficient [15][16][17].…”