InitRech 2015/2016, sujet 12 : Différence entre versions

De Wiki de Projets IMA
(Main contribution)
(Applications)
 
Ligne 33 : Ligne 33 :
  
 
=Applications=
 
=Applications=
 +
 +
It seems that this article can permit applications on the medical domain.
 +
 +
The application introduced in this article is to propose an efficient model able to simulate the final rendering of  fat filling surgery, taking into account the facial anatomy of the considered patient.
 +
 +
It is a pretty specific application, and it is hard to imagine other possible ways to use this simulation model.

Version actuelle datée du 19 juin 2016 à 19:46

Summary

This article deals with a specific reconstructive facial surgery method, called fat-filling. The main purpose of this method is to inject some fat between different layers of tissus, which compose the skin and the muscles of the face, in order to reinflate the face and to fill hemifacial deformations which can be caused by accidents or more often by diseases like the Parry-Romberg Syndrome.

The fat-filling method is kinda different compared with all the plastic surgeries that we hear about every day. Those surgeries, like the breast augmentation or other facial modifications, often use implants, and it is easy to find works ans documentations about them.

The main aim of this research is to propose an efficient model, able to simulate the final rendering of that kind of surgery and the problems that could be encountered. In fact, this model takes into account specific information about the patient, particularly his anatomy (ligaments, nerves and blood vessels), which, in certain cases, can prove to be obstacles for the injected fat.

With those specific datas, and relying on both fluid and deformable solid models, searcher of the INRIA laboratory were able to make their own model. Its aim is to calculate the error which could occur during the projection of the fat, meaning that input fluxes, which have theoretically to be equal to output ones, can differ.

finally, the equation ruling this simulation is the following :

(J+R)p=d
with : d --> vector including divergence values (error named before)
       p --> vector including pressure values
       J --> assembling matrix including area coefficients of the face
       R --> diagonal regularisation matrix summing the contribution of volume displacement by a pressure in the cell n°i over N cells

Some experiments were done :

The first was to simulate a fluid injection between two layers, one was a deformable solid, the other a rigid solid. The conclusion of this test was that the model was coherent with fluid general behavior.

The second was to simulate the surgery with a real patient datas acquired by a laser scan of his face. The aim here was to compare simulation result with the result obtained after a real fat-filling surgery. Finally, the deformations were similar, and that permitted to say that this model was realisitic and efficient, even if there was some differences between the simulation and the surgery outcome.

Main contribution

This article has the objective to propose an efficient way to simulate the result of a fat filling facial surgery. The final aim is to get some datas, sampled over a post-operative patient, and to compare the final rendering of the simulation with the real patient one.

The main contribution of this article is a model created by the combination of two well known models. Those are the fluid model, usefull to predict the behaviour of the fat beneath the skin, and the deformable solid model, here to try to predict the skin deformation during and after the surgery.

These two models are coupled using the computation of the mechanical compliance matrix

Applications

It seems that this article can permit applications on the medical domain.

The application introduced in this article is to propose an efficient model able to simulate the final rendering of fat filling surgery, taking into account the facial anatomy of the considered patient.

It is a pretty specific application, and it is hard to imagine other possible ways to use this simulation model.