MAMCA 2011 - Workshop on Multimodal Audio-based Multimedia Content Analysis (MAMCA 2011)

By definition, multimedia content is composed of multiple forms, including audio, video, text/ subtitles, and others. Traditionally, applications and algorithms that work with such content have considered only a single modality, allowing for example searching of textual tags, thereby ignoring any information available from others modalities. The limitations of this approach are obvious, and there is a recent trend towards multimodal processing, in which different content modalities complement each other, or are used for bootstrapping analysis of new modalities.

Audio is a prominent part of multimedia content, which is backed up by extensive research by the speech and music communities, although usually performed on audio-only systems. Utility of audio-only systems is often limited by the quality of the acoustic environment or the information contained therein, so they can benefit from a multimodal analysis of multimedia data, to enhance the resulting performance, robustness, and efficiency.

The main goal of the workshop is to explore ways in which audio processing can be enhanced, bootstrapped, or facilitated by other available information modalities. We are interested not only in applications that show successful combinations of audio and other sources of information, but also on algorithms that effectively integrate them and leverage complementary information from each modality to obtain an enhanced result, in terms of degree of detail, coverage of the corpus, or other enabling factors.

The workshop will provide a forum for publication of high-quality, novel research on multimedia applications and multimodal processing, with a special focus on the audio modality.

Paper format

MAMCA-2011 solicits regular technical papers of up to 6 pages following the ICME author guidelines. The proceedings of the workshop will be published as part of the IEEE ICME 2011 main conference proceedings and will be indexed by IEEE Xplore. Papers must be original and not submitted to or accepted by any other conference or journal.

Paper submission

Papers can be submitted through https://cmt.research.microsoft.com/ICMEW2011

Selection process

Papers submitted to the workshop will be peer-reviewed by members of the community with extensive experience both in audio processing as well as other relevant modalities considered. The review will be semi-blind and assignment will be performed manually in order to generally produce three best practice reviews of each of the submitted papers.

List of topics

Topics include, but are not limited to:

Effective fusion of audio with other modalities

Multimodal input applications, where one input is audio

Multimodal databases

Bootstrapping of multimodal systems

Co-training for labeling new data

User-in-the loop calculations to detect preferences

Games with a purpose to label new data

Improving robustness through multimodality

Prediction of modality preference

Applications that utilize multimodality

Important dates

Paper submission deadline: February 20th 2011

Paper acceptance notification: April 10th 2011

Camera-ready paper: April 20th 2011

Workshop day: July 11th 2011