Research Projets

Current projets 4

ANR WoodSeer (2019-)

  • Project: Predicting Inner Wood Defects from Outer Bark Features
  • Funding: French Agence Nationale de le Recherche (ANR-19-CE10-011)
  • Proposal: From a biological point of view, similarly to wood recording events occurring during a tree development, the tree bark reflects also its history and particularly the development of its branching. From a living branch until a knot in duramen, these different stages have a more or less evident impact on the external bark roughness, (slight colour differences can also be present). The leading assumption of WoodSeer is that it is possible to characterize automatically surface defects and to link them to the characteristics of the inner part through the establishment of relationships between both.
    From an industrial point of view, WoodSeer targets the supply chain of wood for the first transformation industry but also the first steps of this transformation by aiming at adding information about the quality of the raw material at different steps of the supply processes: for standing trees in the forest, for felled logs stored at the roadside ,or just before their transformation in the sawmill. WoodSeer consortium believes that such information, added to a traceability framework, could be valuable to the wood supply chain, and in particular for the trading processes, adding transparency and offering the possibility to finely assign logs to a specific usage for customers.

ANR TreeTrace (2016-)

  • Project: Biometric fingerprints of trees: log tracing from forest to sawmill and early estimation of wood quality
  • Funding: French Agence Nationale de le Recherche (ANR-17-CE10-0016)
  • Proposal: With the increasing amount of imaging devices installed at sawmills, the importance of using these data for improving workflow and for increasing revenues in the wood processing industries is growing. In this context, challenging questions with respect to imaging and image processing technology arise, several of which will be tackled in this joint project.
    The project considers two application cases as follows: The first application case is the question of tracing tree logs from the forest harvesting site to the sawmill by using biometrics related tree log recognition techniques based on image processing of cross-section data only. This approach of course assumes the additional availability of imaging sensors in the forest. Since there is a trend for installing CT imaging devices at sawmills, which are of course not available in the forest, the challenging issue of cross modality matching arises. The second application case is the determination of wood quality from cross-section imagery, applicable already in the forest, and/or at the sawmill.

ANR R-Vessel-X (2018-)

  • Project: Vascular network extraction and understanding within biomedical images
  • Funding: French Agence Nationale de le Recherche (ANR-18-CE45-0018)
  • Proposal: Cardiovascular diseases and other blood vessels disorders increase in the world-wide scale and in particular in Occident. The evolution of computer science in researches investigating vascular networks has raised the interest in numerical reconstructing and understanding of those complex tree-like structures. The R-VESSEL-X project proposes original and robust developments of image analysis and machine learning algorithms integrating strong mathematical frameworks, e.g. digital geometry and topology, mathematical morphology, or graphs for reconstructing vessels of the liver beyong medical image content. Another objective of R-VESSEL-X is to diffuse research works in an open-source way, with the developments of plug-ins compatible with the ITK and VTK librairies largely popularized by the KITWARE company. This project will also include benchmarks composed of images, associated ground-truth and quality metrics, so that researchers and engineers evaluate their novel contributions.

SolHoM-Fossard (2019-)

  • Project: Interactions hommes-milieux et évolution des sols au cours des deux derniers millénaires dans le massif du Fossard (Remiremont, Vosges)
  • Funding: Project interdisciplinaire de l’Université de Lorraine
  • Proposal: Notre équipe de pédologues, archéologues, géographes et informaticiens s’intéresse aux archives pédo- sédimentaires du massif du Fossard, occupé par l’Homme sur presque 2000 ans. Des transects, sélectionnés à partir de cartes et d’analyses d’image LiDAR, seront décrits et échantillonnés. Les sols seront analysés (physico-chimie, micromorphologie, caractérisation de la matière organique) et datés. L’approche intégrée multi-scalaire et interdisciplinaire permettra de mettre en évidence l’empreinte des activités humaines historiques sur les sols, exacerbées ou non par certains évènements climatiques majeurs.

Past projets 2

Invariance géométrique (April 2017)

  • Project: Invariance géométrique par transformations rigides digitales
  • Funding: bourse de mobilité du GdR IG-RV 2017
  • Proposal: Avec l’explosion des données digitales dans lesquelles les objets sont discrétisés/numérisés, nous nous intéressons à l’utilisation des transformations géométriques, en particulier les transformations rigides, dans le contexte des images numériques. Plus précisément, les transformations appliquées aux images numériques de Z2 sont appelées les transformations rigides digitales. Il s’agit des transformations rigides de R2 qui suivent un procédé de digitalisation afin d’obtenir un résultat sur Z2. Il est montré que ces transformations digitales perdent, dans la plupart des cas, leurs propriétés géométriques par rapport à leurs homologues continus, à cause de la digitalisation.
    Dans ce contexte, nous voulons étudier aussi les conditions/caractérisations des images numériques qui permettent de préserver leurs propriétés géométriques, en particulier la convexité, sous des transformations rigides arbitraires.

Image Matching (January-Mars 2013)

  • Project: 2D Image Matching using Local Search
  • Funding: Université de Paris-Est
  • Proposal: 2D Image Matching (2D-IM, for short) refers to the problem of finding an admissible transformation between two given images. In particular, we address 2D-IM under rigid transformations, i.e., that combine translations and rotations. In a recent work, we proposed to study rigid transformations for digital images as a fully discrete process using a combinatorial structure -namely, a graph- to model the whole space of rigid transformations on arbitrary subset of Z2 of size N × N.
    In this project, we plan to investigate the 2D-IM problem using combinatorial optimisation applied on the proposed structure.