MUDIMO - MUltimodal DIgital MOnitoring
of PLIN and PCOM research
Financing funds : IL&C
The main objective of the MUDIMO project is to develop a set of software tools allowing one-time and/or continuous acquisition (collection and storage), analysis and enrichment, and navigation (search) within large multimodal data spaces. By multimodal data, we mean any corpus of data including at least two of the following modalities: text (multilingual), image, sound, video.
Based on the expertise and tools developed at IL&C in past or ongoing research projects, and based on the needs identified to date, the specific main functionalities that will be implemented within the framework of MUDIMO are as follows: :
- Acquisition: occasional and/or continuous collection of multimodal data, using scraping techniques.
- Enrichment: image analysis (text extraction, object identification, automatic description of scenes) and text analysis (keyword extraction, named entity recognition, topic modeling, sentiment analysis)
- Navigation: search by keywords within different modalities.
