1,720,975 research outputs found
A Video Indexing Approach Based on Audio Classification
No full textThis paper presents a video indexing approach based only on audio classification. Indeed, we apply to an audio-visual document a set of methods for partitioning the associated audio data into homogeneous segments. The aim is to highlight semantically relevant items of a multimedia document by relying only on simple audio processing techniques. A simple algorithm to identify audio segments belonging to silence, music, speech and noise classes has been proposed
Report on Ordering Key DS CE
No full textIn this document, we report the result of our core experiment on the Ordering Key DS. In the last meeting a decision is made to specialize the Weight DS [m6227] in a new DS in order to achieve ordering functionalities. This new DS is the Ordering Key DS which indicates which descriptors can be used in order to obtain an ordering of generic DSs (Segment, Concept, Object, Event). The main objectives of this CE are: an improvement of the syntax which is currently part of the XM document; to clarify explicitly the semantics of this DS so that users can understand what they should obey in their application; to demonstrate the benefit to have such component for ordering purposes
Proposals for OrderingKey DS instantiation in MPEG-7 description
No full textThis document identifies some solutions to the problem stated in the Editor’s Note in clause 7.7.1.1 of the MDS Study of CD, document ISO/IEC JTC 1/SC 29/WG 11/N3816. Therefore this document is intended as a contribution to the MDS group.
Presently there is no way to instantiate an Ordering Key DS within a “valid” MPEG-7 description.
To solve this problem three proposals are here considered and their respective implication are discussed. The first proposal is to derive the OrderingKeyType as an extension of the HeaderType allowing the instantiation, at the header level, of any MPEG-7 Complete/Unit Description and DSType instances. The second solution is to include the OrderingKey DS as a component of the BasicDescriptionType. This allows to apply the ordering tool only in an MPEG-7 Complete Description, but no Unit Description. The third solution is to adopt both the first and second proposals. In Section 2 the three solutions are detailed and the required modification in terms of syntax and usage are discussed. In Section 3 the three solutions are compared
Analysis of Video Content for a Multi-Layer Navigation of Multimedia Documents
No full textThis paper describes a set of automatic extraction tools so as to generate a three-layer organization of video documents. The underlying coarse to fine description allows for a fast navigation throughout the document, depending on the degree of details which is desired. Once the time-codes of the individual segments for each layer of the hierarchy have been identified, it is possible to map them into a Description Scheme (DS), which maintains the hierarchy and linear structure of the video document. This structural DS serves the role of a table of content for the multimedia document, the same way it is done in books. The particular interest of the proposed approach lies in the automatic solutions that can be used to generate the different segments at each level of the DS, and in the browsing tool that can be easily derived to navigate throughout the document
Indicizzazione di sequenze video attraverso l'analisi dell'audio sottostante
No full textIn questo lavoro viene proposto un approccio innovativo all'indicizzazione di sequenze video basato sull'analisi dello stream audio sottostante. I metodi implementati permettono di segmentare l'audio in un insieme di classi immagine con elevato contenuto semantico. Ciò consente di individuare situazioni altamente significative solamente informazioni estratte attraverso l'elaborazione del segnale audio. Per far questo è stato sviluppato un algoritmo di segmentazione dell'audio in quattro classi omogenee (silenzio, voce, musica, rumore) basato sull'analisi di semplici caratteristiche temporali e frequenziali
TOCAI: uno schema di descrizione per l'indicizzazione ed il recupero di documenti multimediali
No full textIn questo articolo viene presentato uno schema di descrizione DS (Description Scheme) denominato con l'acronimo ToCAI (Table of Content-Analytical Index). L'idea di fondo è di sfruttare una modalità simile a quella adottata per capire l'organizzazione ed il contenuto di un libro tecnico per esplorare un documento Audio-Video (AV). Dall'indice ToC (Table of Contents) è possibile capire l'organizzazione sequenziale in cui sono ordinati i diversi capitoli mentre tramite l'indice analitico AI (Analytical Index) è possibile arrivare in modo veloce ad argomenti o termini di interesse. L'utilizzo di questo tipo di schema permette una rappresentazione gerarchica dell'organizzazione temporale di un documento AV, (tramite la sezione ToC) molto utile per esplorare in modo veloce, assieme ad un indice analitico (AI) che punta ad oggetti multimediali di interesse. Vengono esaminati anche due sotto schemi di metadata associati al documento. Viene descritta la struttura dettagliata del DS utilizzando la notazione UML per passare poi ad un esempio reale. Infine sono state riportate alcune considerazioni sull'interfaccia realizzata
The TOCAI Description Scheme for Indexing and Retrieval Multimedia Documents
No full textA framework, called Table of Content-Analytical Index (ToCAI), for the content description of multimedia
material is presented. The idea for such a description scheme (DS) comes out from the structures used for
indexing technical books (containing a Table of Content, typically placed at the beginning of the book, where the
list of topics is organized hierarchically into chapters, sections, and an Analytical Index, typically placed at the
end of the book, where keywords are listed alphabetically). The ToCAI description scheme provides similarly a
hierarchical description of the time sequential structure of a multimedia document (ToC), suitable for browsing,
and an “Analytical Index” (AI) of audio-visual key items for the document, suitable for effective retrieval. Besides
two other sub-description schemes are proposed to specify the program category and the description of other
metadata associated to the multimedia document in the general DS. The detailed structure of the DS is presented
by means of a UML diagram. Moreover, some suitable automatic extraction methods for the identification of the
values associated to the descriptors that compose the ToCAI are presented and discussed. Finally, a browsing
application example is also proposed
Low Level Processing of Audio and Video Information for Extracting the Semantics of Content
No full textThe problem of semantic indexing of multimedia documents is actually of great interest due to the wide diffusion of large audio-video databases. We first briefly describe some techniques used to extract low-level features (e.g., shot change detection, dominant color extraction, audio classification etc.). Then the ToCAI (table of contents and analytical index) framework for content description of multimedia material is presented, together with an application which implements it. Finally we propose two algorithms suitable for extracting the high level semantics of a multimedia document. The first is based on finite-state machines and low-level motion indices, whereas the second uses hidden Markov models
Report on Ordering Key DS CE
No full textIn this document, we report the results of our core experiment on the Ordering Key D. The concept of the Ordering Key D has been proposed for the first time in Beijing meeting [m6227] as a part of the generic Weight DS. According to the discussions during that meeting the Weight DS has been specialized in order to obtain exclusively ordering functionalities [w3473]. At the same time the demo showed in Beijing seemed to demonstrate the useful of the ordering beetwen entities of various types. At the following meeting a specialized and improved semantic and syntax have been proposed. In spite of positive results about its utility, as shown in the demo [m6588] and also in the document [m6491], it was not clear the use of XPath syntax as mechanism to obtain reference to entities. At the La Baule meeting the DDL group had not yet decided about the use of XPath as reference datatype. In conclusion the aim of this CE was not to demonstrate the OrderingKey effectiveness but to investigate about its syntax [w3630]
Validation Experiment on the Ordering Key DS and an Unified Syntax for the Weight DS
No full textThis document presents the experimental results for validating the Ordering Key DS (DS5) in the context of the core experiment of the Weight DS [6]. At the Melbourne MPEG meeting, in October 1999, the aforementioned core/validation experiment was planned in order to show the validity of a set of proposals (Weight DS [5], Descriptor Usage DS [8], Fidelity DS [9], Pointofview DS [4] and Ordering Key DS [7]). In a few words, all these DSs play the role of highlighting, by means of some kinds of weights, description information (DSs or Ds) relevant to user queries. They can provide, for example, confidence measure, priority, fidelity, relevance feedback, information for ordering etc. in order to facilitate user queries and browsing. As we said, the document focuses on the VE of the Ordering Key DS. Besides it presents an unified DS, in MPEG-7 DDL syntax, that addresses all the different functionalities proposed by the several DSs involved in the CE. In our case, the provided functionality deals with the concept of ordering, as we consider the need of providing users with ordering mechanisms a very relevant issue for MPEG-7. Such ordering mechanisms are derivable from descriptors (e.g. a set of key – frames ordered on the basis a color descriptor or a set of sounds ordered by means of an audio loudness D). However a possible large variety in the types of descriptors composing a description could lead to a consequent high number of ordering criteria to arrange description items. Therefore we propose that the description provider (it could be different from the content provider) should also select a reduced set of descriptors allowing to order a subset of description elements (e.g. key frames, events etc.) pertinent to the MM document being described [7]. This contribution is organized as follows. In Section 2, the motivations for introducing Ordering Key DS in the Generic AV DS are given. Section 3 briefly presents the structure of the Ordering Key DS by means of UML notation and MPEG-7 DDL as well. Moreover the section discusses about the possible locations of the DS within the Generic AV DS. In Section 4 is shown the output of the experiment: a smart browsing of ordered elements belonging to the test set of MPEG-7 video material. Finally, in Section 5, is proposed a unified DDL structure for the DSs involved in the CE
- …
