Region growing is one of the most popular image segmentation methods. The algorithm for region growing is easily understandable but criticized for its lack of theoretical background. In order to overcome this weakness, we propose to
describe region growing in a new framework using a variational approach that we called Variational Region Growing
(VRG). Variational approach is commonly used in image segmentation methods such as active contours or level sets, but is rather original in the context of region growing. It relies on an evolution equation derived from an energy minimization, that drives the evolving region towards the targeted solution. Here, the energy minimization and the VRG robustness to the initial seeds location are performed on gray-level and color images. (pdf)

Jean-Loïc Rose(a), Thomas Grenier(b), Chantal Revol-Muller(b), Christophe Odet(b)
(a)LIRIS, CNRS UMR5205, Université Lyon 2, 69676 Bron cedex, France
(b)CREATIS, CNRS UMR5220, Inserm U630, Université de Lyon, INSA de Lyon, France

by D. Charpigny, J-C. Brisset, T. Grenier, M. Wiart, and H. Benoit-Cattin

UltraSmall SuperParamagnetic Iron Oxide (USPIO) particles are used in MRI contrast agents for diagnosing different pathologies such as stroke and cancer. Determining the concentration of USPIO in MRI is of great interest. Here we present a non invasive quantification process of the USPIOs’ concentration from MR images based on the physical effect of these nanoparticles induced by the difference of magnetic susceptibility. pdf (149kB)

3D Robust Adaptive Region Growing for segmenting [¹⁸F]fluoride ion PET images

We propose a new Robust Adaptive Region Growing method (RoAd RG) based on two local parameters: the local mean value of the intensity function and the local mean value of the norm of the intensity gradient. This approach enables a better spread of the region growing inside the region of interest while avoiding the merge of outlier pixels. We applied positively our method to 3D [18F]fluoride ion PET images for segmenting bone structures and showed its superiority compared to a non adaptive method. pdf (336kB)

T. Grenier, C. Revol-Muller, N. Costes, M. Janier, G. Gimenez.

Multiparametric smoothing based on Mean shift procedure for ultrasound data segmentation

Segmentation of ultrasound data is improved when using multi-parametric approach. In this paper we propose the use of Multi-Parametric Mean Shift procedure (MPMS). Two derived processes are described: MPMS smoothing which achieves a multi-parametric filtering in the spatial-range domain and MPMS segmentation which takes benefit of this filtering for segmenting the multidimensional data. MPMS segmentation is particularly attractive, since it achieves an unsupervised segmentation. These methods were positively tested on three sets of simulated ultrasonic data, representative of various scatterers densities and also various scattering conditions. pdf (690kB) Poster(733kB)

Thomas Grenier, Chantal Revol-Muller, Franck Davignon, Olivier Basset, Gérard Gimenez.

Variable bandwidth Mean Shift for Smoothing ultrasonic images

As the variance of the statistics of ultrasonic data in a homogeneous tissue may be rather large and the statistics of different tissues may be very similar, a new filtering approach is proposed to enhance the contrast in ultrasonic images. It is based on the Variable Bandwidth Mean Shift algorithm adapted to the specificities of ultrasonic data. A fully automatic adaptive bandwidth selection in both range and spatial domains is described. Our method was compared to a Variable Bandwidth Mean Shift algorithm based on an adaptive range scale selection and a fixed spatial scale parameter. The results show the superiority of our method. pdf (410kB) Poster (176kB)

Thomas Grenier, Chantal Revol-Muller, Franck Davignon, Olivier Basset, Gérard Gimenez.

Mean shift