TOPACS

Purpose and context

The ANR-funded TOPACS project proposes to investigate full-body computational anatomy, through the analysis of a large set of 3D CT and MRI scans (more than 10000 individuals) stored in Picture Archiving and Communication Systems (PACS) of partner hospitals. As 3D scans provide volumetric information, it is possible to take into account inner organs and gather significantly more information than by considering only the outer body envelope. First applications will focus on Forensic Science, such as the determination of age and sex of unknown individuals.

Project Members

CREATIS

• Sébastien Valette (Project Leader)
• Razmig Kéchichian
• Laurent Fanton

Institut Pascal

• Adrien Bartoli

ICube

• Fabrice Heitz
• Vincent Noblet
• Sylvain Faisan

Gipsa-Lab

• Michel Desvignes

LIRIS

• Raphaëlle Chaine
• Julie Digne

CURML

• Tony Fracasso

Publications resulting from the project

• Sex estimation from coxal bones using deep learning in a population balanced by sex and age

  In the field of forensic anthropology, researchers aim to identify anonymous human remains and determine the cause and circumstances of …

  Marie Epain, Sébastien Valette, Kaifeng Zou, Sylvain Faisan, Fabrice Heitz, Pierre Croisille, Tony Fracasso, Laurent Fanton

  Cite DOI

  
• 4D Facial Expression Diffusion Model

  Facial expression generation is one of the most challenging and long-sought aspects of character animation, with many interesting …

  Kaifeng Zou, Sylvain Faisan, Boyang Yu, Sébastien Valette, Hyewon Seo

  PDF Cite DOI

  
• Disentangling high-level factors and their features with Conditional Vector Quantized VAEs

  Two recent works have shown the benefit of modeling both high-level factors and their related features to learn disentangled …

  Kaifeng Zou, Sylvain Faisan, Fabrice Heitz, Sébastien Valette

  PDF Cite DOI

  
• Disentangled representations: towards interpretation of sex determination from hip bone

  Neural network-based classification methods are often criticized for their lack of interpretability and explainability. By highlighting …

  Kaifeng Zou, Sylvain Faisan, Fabrice Heitz, Marie Epain, Pierre Croisille, Laurent Fanton, Sébastien Valette

  PDF Cite DOI

  
• Joint disentanglement of labels and their features with VAE

  Most of previous semi-supervised methods that seek to obtain disentangled representations using variational autoencoders divide the …

  Kaifeng Zou, Sylvain Faisan, Fabrice Heitz, Sébastien Valette

  PDF Cite DOI

  

• Procrustes Analysis with Deformations: A Closed-Form Solution by Eigenvalue Decomposition

  International Journal of Computer Vision
  Fang Bai, Adrien Bartoli
  PDF DOI
• Learning 3D Medical Image Keypoint Descriptors with the Triplet Loss

  Purpose: We propose to learn a 3D keypoint descriptor which we use to match keypoints extracted from full-body CT scans. Our methods …

  Nicolas Loiseau--Witon, Razmig Kéchichian, Sébastien Valette, Adrien Bartoli

  PDF Cite DOI

  
• Local SURF-based Keypoint Transfer Segmentation

  This paper presents an improvement of the keypoint transfer method for the segmentation of 3D medical images. Our approach is based on …

  Antoine Bralet, Razmig Kéchichian, Sébastien Valette

  PDF Cite DOI

  
• Learning 3D Medical Image Patch Descriptors with the Triplet Loss

  Computational anatomy focuses on the analysis of the human anatomical vari- ability. Typical applications are the discovery of …

  Nicolas Loiseau--Witon, Razmig Kéchichian, Sébastien Valette, Adrien Bartoli

  PDF Cite