Database Availability. The data to be produced in this WP has the potential to be extremely useful in future research in multimodal processing. The consortium is willing, in principle, to make this data available for use to the research community. To this end, we shall open negotiations with the Linguistic Data Consortium to discuss this possibility. To make these databases publicly available, several issues must be resolved in particular relating to privacy and logistics of distribution (the full M4 multistream database will probably occupy tens or hundreds of gigabytes). WP2: Multimodal Recognition WP Overview This WP is concerned with the development, evaluation and improvement of the different input modal- ities that will be integrated in WP3. This includes robust conversational speech recognition, gesture and action recognition, emotion/intent recognition, source localization and tracking, and multimodal person identification and tracking. While different versions of these components are already available in the M4 consortium, significant progress can be expected from the present project. Although different, a key bind- ing factor is that these technologies are often based on similar or complementary mathematical tools, thus providing good opportunities for collaboration.
Appears in 2 contracts
Samples: spandh.dcs.shef.ac.uk, spandh.dcs.shef.ac.uk
Database Availability. The data to be produced in this WP has the potential to be extremely useful in future research in multimodal processing. The consortium is willing, in principle, to make this data available for use to the research community. To this end, we shall open negotiations with the Linguistic Data Consortium to discuss this possibility. To make these databases publicly available, several issues must be resolved in particular relating to privacy and logistics of distribution (the full M4 multistream database will probably occupy tens or hundreds of gigabytes). WP2: Multimodal Recognition MULTIMODAL RECOGNITION WP Overview This WP is concerned with the development, evaluation and improvement of the different input modal- ities that will be integrated in WP3. This includes robust conversational speech recognition, gesture and action recognition, emotion/intent recognition, source localization and tracking, and multimodal person identification and tracking. While different versions of these components are already available in the M4 consortium, significant progress can be expected from the present project. Although different, a key bind- ing factor is that these technologies are often based on similar or complementary mathematical tools, thus providing good opportunities for collaboration.
Appears in 2 contracts
Samples: spandh.dcs.shef.ac.uk, spandh.dcs.shef.ac.uk