SIROCCO Research team
Analysis representation, compression and communication of visual data
- Leader : Christine Guillemot
- Type : Project team
- Research center(s) : Rennes
- Field : Perception, Cognition and Interaction
- Theme : Vision, perception and multimedia interpretation
- Partner(s) : CNRS,Université Rennes 1
- Collaborator(s) : Institut de recherche en informatique et systèmes aléatoires (IRISA) (UMR6074)
Efficient processing, i.e. analysis, storage, access and transmission of visual content, with continuously increasing raw data rates, in environments which are more and more mobile and distributed, remains a key challenge of the years to come. The emergence of new image modalities leads to a sustained need for algorithmic tools allowing efficient compression and communication of large volumes of visual data, of visual features and descriptors extracted for different video processing tasks.
The goal of the project-team is the design of algorithms and practical solutions in the areas of visual data analysis, modeling, representation, compression and communication. Our activities are thus structured around the following inter-dependent axes:
- Analysis and modeling for compact representation and navigation in large volumes of visual data
- Representation and compression of visual data
- Distributed processing and robust communication of visual data
The proposed research is at the frontier of computer vision, signal processing, coding and information theory. In terms of application domains, the project will primarily target networked visual applications such as 3DTV, FTV, camera sensor networks, satellite and medical imaging applications.
Our research activities are structured into three main research axis:
- Analysis and modeling for compact representation and navigation in large volumes of visual data. This research axis includes the extraction of different visual cues such as scene geometry, epitomes, or other high-level features, as well as the study of computational models of visual attention, which are useful for different visual processing tasks.
- Representation and compression of large volumes of visual data. This research activity concerns the development algorithmic tools for constructing low-dimensional and sparse representations of various types of visual data (2D images and video, 3D and multi-view data). The design of sparse representation tools requires tackling other issues such as dictionary learning.
- Distributed processing and robust communication of visual data. The goal is to develop theoretical and practical solutions for robust image and video transmission over heterogeneous and time-varying networks. This includes the design of network aware source coding solutions, of loss concealment algorithms and of data protection schemes which would be robust to channel uncertainties. Source aware network coding, and distributed coding solutions are also studied.
International and industrial relations
The project-team has a number of industrial collaborations via CIFRE contracts with Orange, Technicolor, Thomson Video Networks and bi-lateral collaborations with Astrium, with Alcatel Lucent in the context of the joint INRIA/Alcatel lab. The project-team also contributes to the MPEG-3DV standardization.
Research teams of the same theme :
- LINKMEDIA - Creating and exploiting explicit links between multimedia fragments
- MAGRIT - Visual Augmentation of Complex Environments
- MORPHEO - Capture and Analysis of Shapes in Motion
- PERCEPTION - Interpretation and Modelling of Images and Videos
- STARS - Spatio-Temporal Activity Recognition Systems
- THOTH - Learning visual models from large-scale data
- WILLOW - Models of visual object recognition and scene understanding
This team follows
Search for a team
By Inria research center