Kameleon : Masse de Données Anatomiques pour L’Étude des Structures Squelettiques des Petits Vertebras

Kameleon : Masse de Données Anatomiques pour L’Étude des Structures Squelettiques des Petits Vertebras

Lionel Revéret Estelle Duveau  Marc Herbin  Franck Multon  Pierre-Paul Vidal 

Universités Grenoble & CNRS, LJK (UMR 5224), INRIA Grenoble Rhône-Alpes

Universités Grenoble & CNRS, LJK (UMR 5224), INRIA Grenoble Rhône-Alpes Département d’Ecologie et Gestion de la Biodiversité (CNRS UMR 7179), Muséum national d’histoire naturelle

Laboratoire Mouvement Sport Santé, Université Rennes

Centre d’Etude de la SensoriMotricité (CNRS UMR 8194), Université Paris Descartes

Page: 
309-328
|
DOI: 
https://doi.org/10.3166/TS.28.309-328
Received: 
N/A
|
Accepted: 
N/A
|
Published: 
30 June 2011
| Citation

OPEN ACCESS

Abstract: 

Measurement of motion for small vertebrates such as mice and rats is a key challenge in biology and clinical research. This measurement is very complex due to the inherent structure of these animals. Although skeletal information would be the most deterministic information, this information is hidden within the animals’ body and individual elements are difficult to identify. Through the ANR project Kameleon, we describe in this paper an experimental platform for the three dimensional study of small vertebrates. The overall approach of this platform is to combine a multiple view camera system for 3D measurement by stereoscopy with an Xray video system imaging the skeletal structure in motion. We present here the methodology for the set up of this platform and some results.

Extended Abstract

Rodents are largely used as animal model in biology and clinical research. In this context, motion analysis systems are classically restricted to the 2D location of the animal in the ground plane. However, in neurophysiology of motor control, the movement of 3D skeletal structure is carrying important information which should be considered for measurement. In medical research, such a new source of biometrics measurement would bring new way to diagnose neuromotor disorder. In computer graphics, the visualization of complex structure of animal in motion is an interesting challenge. Traditional motion capture techniques do not allow these requirements. In this paper, we describe a new platform for the 3D measurement of motion of small vertebrates. The overall approach of this platform is to combine a multiple view camera system for 3D measurement by stereoscopy with an Xray video system imaging the skeletal structure in motion. We present here the methodology for the set up of this platform and some results. We firstly describe the tentative to apply existing techniques in 3D reconstruction to our specific experimental condition. In particular, we emphasis the difficulties related to rodent where, contrary to humans, fur, fat and muscles make impossible an easy retrieval of the skeletal structure. In order to collect 3D information, the volume space needs to be geometrically calibrated. We describe our methodology to calibrate both the multiple camera video system for the measurement of the external surface, and the Xray video system for the internal skeletal structure. Once calibrated, the set up allows to record 3D information. We first validated the ability of the system to retrieve 3D trajectories of markers attached to the animal’s external anatomical landmarks. We have achieved an accuracy of less than 1mm at a sampling frequency of 200Hz. Secondly, we report experiments for the reconstruction of 3D surface of the whole body in motion. While results are promising, it turned out that additional geometry processing is required to obtain a smooth and complete surface. In parallel to the measurement of an external surface, we have developed anatomical processing of the internal view of the skeletal structure. Firstly, we have build off-line an accurate 3D modeling of the skeleton of a rat. We had access to the ESRF synchrotron facility in Grenoble to collect a 3D volumetric image of the skeleton with an accuracy of 45 μm. Based on these data, a biomechanical model of a rat skeleton has been developed with articulated joints. An Inverse-Kinematic method has been developed to drive this biomechanical model from the 3D measurement, points and surfaces, obtained from the previous video analysis of external views. To do so, angular springs are located at the articulated joint. The Inverse-Kinematic system consists in the optimization of an energy term, combining the preference for a rest post and an attraction to 3D motion data. The preference for the rest pose is modeled as the angular spring potential energy. The attraction to motion data is modeled with the potential energy of linear springs,attached at specific location of the bones and to their closest location on the external surface, measured in term of points or surface. In addition to the pilot study on rat, this platform has been used for studies on other animals such as lizards and lemurs.

RÉSUMÉ

La mesure de mouvement des petits vertébrés tels que souris et rats représente un enjeu important en biologie et recherche clinique. Elle est cependant rendue très complexe de par la structure naturelle de ces animaux. Alors que l’information squelettique serait la plus à même de donner une information non ambiguë sur la posture de l’animal, celle-ci est enfouie, ne permettant pas une identification simple des éléments la constituant. A travers le projet ANR Kameleon, nous décrivons dans le présent article une plateforme d’étude tridimensionnelle du mouvement des petits vertébrés. Le principe de cette plateforme est de coupler un système multicaméra permettant d’extraire des informations 3D par stéréoscopie avec un système de cinéradiographie par rayon X permettant d’avoir une vue des structures squelettiques en mouvement. Nous présentons ici la démarche suivie pour la construction de cette plateforme, ainsi qu’une partie des résultats.

Keywords: 

anatomy, vertebrates, 3D modelling, 3D animation, motion capture, high-speed video, Xray video.

MOTS-CLÉS

anatomie, vertébrés, modélisation 3D, animation 3D, capture de mouvement, vidéo haute-fréquence, vidéoradiographie.

1. Introduction
2. Plateforme d’acquisition Vidéo
3. Numérisation de Surfaces 3D
4. Numérisation du Squelette
5. Applications
6. Conclusions
  References

Amiri S., Wilson D., Masri B., Sharma G., Anglin C. (2011). « A novel multi-planar radiography method for three dimensional pose reconstruction of the patellofemoral and tibiofemoral joints after arthroplasty. », J Biomech.

Basri R., Jacobs D. (2007). « Photometric stereo with General, Unknown Lighting », International Journal of Computer Vision, vol. 72, n° 3.

Bauman J. M., Chang Y.-H. (2010). « High-speed X-ray video demonstrates significant skin movement errors with standard optical kinematics during rat locomotion », Journal of Neuroscience Methods, vol. 186, n°1, p. 18-24.

Bey M., Zauel R., Brock S., Tashman S. (2006). « Validation of a new model-based tracking technique for measuring three-dimensional, in vivo glenohumeral joint kinematics. », J Biomech Eng, vol. 128, no 4, p. 604-9.

Bey M., Kline S., Zauel R., Lock T., Kolowich P. (2007). « Measuring dynamic in-vivo glenohumeral joint kinematics : Technique and preliminary results. », J Biomech.

Brainerd E., Baier D., Gatesy S., Hedrick T., Metzger K., Gilbert S., Crisco J. (2010). « X-ray reconstruction of moving morphology (XROMM) : precision, accuracy and applications in comparative biomechanics research. », J Exp Zool A Ecol Genet Physiol. 

Couto P., Filipe V., Magalhães L., Pereira J., Costa L., Melo-Pinto P., Bulas-Cruz J., Maurício A., Geuna S., Varejão A. (2008). « A comparison of two-dimensional and threedimensional techniques for the determination of hindlimb kinematics during treadmill locomotion in rats following spinal cord injury. », J Neurosci Methods.

Defrate L., Sun H., Gill T., Rubash H., Li G. (2004). « In vivo tibiofemoral contact analysis using 3D MRI-based knee models. », J Biomech, vol. 37, n° 10, p. 1499-504.

Filipe V. M., Pereira J. E., Costa L. M., Maurício A. C., Couto P. A., Melo-Pinto P., Varejão A. S. (2006). « Effect of skin movement on the analysis of hind-limb kinematics during treadmill locomotion in rats », Journal of Neuroscience Methods, vol. 153, n° 1, p. 55 - 61.

Gatesy S., Baier D., Jenkins F., Dial K. (2010). « Scientific rotoscoping : a morphology-based method of 3-D motion analysis and visualization. », J Exp Zool A Ecol Genet Physiol.

Hartley R. I., Zisserman A. (2004). Multiple View Geometry in Computer Vision, Cambridge University Press, ISBN : 0521540518, second édition.

Hayakawa H. (1994). « Photometric stereo under a light source with arbitrary motion », J. Opt. Soc. Am. A, vol. 11, n° 11, page 3079, OSA.

Hertzmann A., Seitz S. M. (2005). « Example-Based Photometric Stereo : Shape Reconstruction with General, Varying BRDFs », PAMI, p. 1254–1264.

Lorensen W. E., Cline H. E. (1987). « Marching cubes : A high resolution 3D surface construction algorithm », SIGGRAPH Comput. Graph., vol. 21, p. 163–169, ACM.

Montuelle S., Herrel A., Libourel P.-A., Reveret L., Bels V. (2009). « Locomotor-feeding coupling during prey capture in a lizard Gerrhosaurus major : effects on prehension mode », Journal of Experimental Biology, vol. 212, p. 768–777.

Prados E., Faugeras O. (2006). « Shape from shading », Mathematical models of computer vision : the handbook, Springer.

Prins A., Kaptein B., Stoel B., Reiber J., Valstar E. (2009). « Detecting femur-insert collisions to improve precision of fluoroscopic knee arthroplasty analysis. », J Biomech.

Sakoe H. (1978). « Dynamic programming algorithm optimization for spoken word recognition », IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 26, p. 43-49.

Tashman S., Anderst W. (2003). « In-vivo measurement of dynamic joint motion using high speed biplane radiography and CT : application to canine ACL deficiency. », J Biomech Eng, vol. 125, n° 2, p. 238-45.

Vogiatzis G., Favaro P., Cipolla R. (2005). « Using Frontier Points to Recover Shape, Reflectance and Illumunation », Tenth IEEE International Conference on Computer Vision (ICCV), vol. 1, p. 228-235.

Wang J., Blackburn T. J. (2000). « X-ray Image Intensifiers for Fluoroscopy ».

Woodham R. J. (1980). « Photometric method for determining surface orientation from multiple images », Optical Engineering, vol. 19, n° 1, p. 139-144.

Xu Y., Chowdhury A. K. R. (2005). « Integrating the Effects of Motion, Illumination and Structure in Video Sequences », ICCV.

Yuille A. L., Snow D., Epstein R., N. B. P. (1999). « Determining Generative Models of Objects Under Varying Illumination : Shape and Albedo from Multiple Images Using SVD and Integrability », International Journal of Computer Vision, vol. 35, n° 3.

Zhang L., Snavely N., Curless B., Seitz S. M. (2004). « Spacetime faces : high resolution capture for modeling and animation », ACM Trans. Graph., vol. 23, p. 548–558, ACM.

Zordan V. B., Van Der Horst N. C.0 (2003). « Mapping optical motion capture data to skeletal motion using a physical model », Proceedings of the 2003 ACM SIG-GRAPH/Eurographics symposium on Computer animation, SCA ’03, Aire-la-Ville, Switzerland, Switzerland, Eurographics Association, p. 245–250.