Recent and current projects

Visual scene understanding [more]

• Unsupervised domain adaptation for semantic segmentation and object detection
• Multi-sensor fusion
• Semantic 3D reconstruction (coll. with Phil Torr’s group at University of Oxford)

Face modeling [more]

(coll. with Christian Theobalt’s group at MPI)

• Personnalized face rig extraction and tracking
• Photo-real face editing and retargeting
• Fast 3D face models from photos with deep learning

Dynamic visual content analysis [more]

• Visual tracking
• Segmentation of video objects
• Analaysis of visual dynamics (coll. with Patrick Bouthemy at Inria)

Graph signal processing and learning [more]

• Structured sampling on graph
• ConvNets on graphs
• Deep solvers for fast and flexible style transfer

Visual inpainting [more]

• Robust patch-based video inpainting (coll. with Andrés Almansa and Yann Gousseau at Telecom ParisTech)
• Structural image inpainting with context encoders
• Face inpainting

Large scale indexing and sketching [more]

• Beyond product quantization for image indexing (coll. with Joaquin Zepeda at Amazon)
• Sketching for large scale unsupervised learning (coll. with Rémi Gribonval at Inria)
• Cross-modal search with deep text-semantic embeddings (coll. with Matthieu Cord at UPMC)

Audio modeling and processing [more]

• Informed source separation
• Audio inpainting and declipping, speech inpainting
• Deep audio style transfer