Abstract Texts displayed in lecture videos are closely related to the lecture content. Therefore, they provide a valuablesource for indexing and retrieving lecture videos. Textual content can be detected, extracted and analyzed automatically byvideo OCR (Optical Character Recognition) techniques. In this paper, we present an approach for automated lecture videoindexing based on video OCR technology: first, we developed a novel video... Texts displayed in lecture videos are closely related to the lecture content. Therefore, they provide a valuablesource for indexing and retrieving lecture videos. Textual content can be detected, extracted and analyzed automatically byvideo OCR (Optical Character Recognition) techniques. In this paper, we present an approach for automated lecture videoindexing based on video OCR technology: first, we developed a novel video segmenter for the structure analysis of slidevideos. Having adopted a localization and verification scheme, we perform text detection secondly. We apply SWT (StrokeWidth Transform) not only to remove false alarms from the text detection stage, but also to analyze the slide structure further.To recognize texts, a multi-hypotheses framework is applied, that consists of multiple text binarization, OCR, spell checkingand result merging processes. Finally, we have implemented a novel algorithm for extracting lecture structure from the OCR-transcript by using geometrical information and text stroke width of detected text lines. We use both segmented key framesand extracted lecture outlines for the further video indexing. The accuracy of the proposed approach is proven by evaluation. Read More Read Less

Abstract Effective and flexible solutions for enabling reduced resources consumption and handling formats heterogeneityare essential in multimedia management systems. The reduction of the resources consumption for indexing can be achievedin two ways: by limiting the multimedia content transfer over the network, and by employing only the most appropriatealgorithms, only over the relevant content. Multiple formats are developed for the description of... Effective and flexible solutions for enabling reduced resources consumption and handling formats heterogeneityare essential in multimedia management systems. The reduction of the resources consumption for indexing can be achievedin two ways: by limiting the multimedia content transfer over the network, and by employing only the most appropriatealgorithms, only over the relevant content. Multiple formats are developed for the description of resources, contents andapplications, which are not necessary interoperable. A solution for coping with this heterogeneity is to create an integrativemodel for the description of any resources. We present in this paper the solution implemented in the context of the LINDOproject for indexing multimedia content within a distributed system. Such solution addresses these two points by reducing asmuch as possible the resources consumption through (1) a distributed indexing technique that avoids multimedia transfer;(2) a flexible mechanism for selecting the indexing algorithms to be employed on each remote server, according to themultimedia content characteristics, its acquisition context and user queries history. Our proposal is based on genericdescription models associated to multimedia content, indexing algorithms and servers. Read More Read Less

Abstract This article describes a system that is able to perform online sound localization, separation and speaker recognition for the purpose of channel assignment in teleconferencing scenarios. The assigned channels are then virtually synthesized to different positions around the listener by a conference server using Head-Related Transfer Functions (HRTFs), which enables the user to identify active speakers in an audio conference.... This article describes a system that is able to perform online sound localization, separation and speaker recognition for the purpose of channel assignment in teleconferencing scenarios. The assigned channels are then virtually synthesized to different positions around the listener by a conference server using Head-Related Transfer Functions (HRTFs), which enables the user to identify active speakers in an audio conference. Intelligibility of communication is also increased by making use of the cocktail party effect which describes how a user can concentrate on one specific participant, even if there are more simultaneously active speakers. To support regular telephone systems, we focus in particular on compatibility to existing Voice over IP (VoIP) infrastructure by using standardized protocols and central signal processing. Both traditional telephone devices and software clients are put into one heterogeneous conference. While the conference server transmits the down-mixed, single-channel conference signal to regular phones, the software clients can receive a stereosignal enabling full 3D audio. Read More Read Less