Signal, Image, Langage (D6)

The growing complexity, heterogeneity and volume of data lead to an increasing need for tools allowing their analysis and processing, most often automated, for a wide variety of application purposes.

In this context, department D6 is interested in signals, images and language for their capture, analysis, compression, interpretation, recognition and generation.

Thus, the added value of the department teams is their ability to process data of all kinds (in particular handwritten and textual documents, video, 2D and 3D images, complex images, signals from sensors, gesture and speech) by developing and using intensively artificial learning techniques (artificial intelligence, deep learning).

Department 6 teams are complementary from the point of view of the types of data processed and share a common base of artificial intelligence techniques.

The main areas of research covered are:

• Analysis, recognition, representation and interpretation of data (bio-medical images, handwritten documents, visual data, satellite imagery, etc.)
• Processing of large-scale document collections, information retrieval, classification
• Person-Document and Person-Machine Interaction
• Synthesis analysis and recognition of expressive gestures and expressive speech
• Expressiveness in texts, text mining
• Methodology and theory of automatic and deep learning