Showing papers by "Christine Fernandez-Maloigne published in 2002"

Segmentation of multi-modality MR images by means of evidence theory for 3D reconstruction of brain tumors

Abstract: In this paper, we propose a segmentation scheme for magnetic resonance (MR) images based on a two step algorithm. The first step consists of a classification based on an evidential k-NN rule initially proposed by Denoeux (1995). The second step allows to take into account the spatial dependence of each voxel of the MR volume in order to lead the segmentation. The goal is to locate properly tumors in MR images of the brain allowing the 3D reconstruction of the different brain structures and the tumor. It can help clinicians observe the tumors accurately and to follow the evolution of the tumors in multidate acquisitions of MR images.

Color texture recognition and indexing by fuzzy color spatial distributions

Abstract: This paper proposes two color texture descriptors based on the introduction of a fuzzy color symilarity function: the spatial color similarity pattern and the spatial color symmetry spread index. Both descriptors embed information on the spatial color distribution within a first-order statistical distribution. Experiments show that the proposed descriptors can be successfully used for color texture recognition and retrieval.

Psychovisual evaluation of the effect of color spaces and color quantification in JPEG2000 image compression

Abstract: JPEG2000 is an emerging standard for still image compression. It is not only intended to provide rate-distortion and subjective image quality performance superior to existing standards, but also to provide features and additional functionalities that current standards can not address sufficiently such as lossless and lossy compression, progressive transmission by pixel accuracy and by resolution, etc. Currently the JPEG2000 standard is set up for use with the sRGB three-component color space.the aim of this research is to determine thanks to psychovisual experiences whether or not the color space selected will significantly improve the image compression. The RGB, XYZ, CIELAB, CIELUV, YIQ, YCrCb and YUV color spaces were examined and compared. In addition, we started a psychovisual evaluation on the effect of color quantification on JPEG2000 image compression. the final results indicate that the YCrCb color space is very suitable for this standard as mentioned in the amendement2.

Nouvelle utilisation de l’information spatiale pour l’indexation d’images basée contenu

Abstract: En indexation d’images, la plupart des descripteurs developpes sont bases sur des statistiques globales des images. Aujourd’hui, des methodes essayant d’aller plus loin dans la representation apparaissent. Sans aller jusqu ’aune description semantique, elles representent les images plus finement en utilisant des statistiques localisees ainsi que les interactions entre celles-ci. Dans la plupart des cas, des structures de graphes sont utilisees. Cet article presente ces diverses approches ainsi que celle que nous avons developpee en construisant un graphe pyramidal. Les premiers resultats obtenus pour des requetes globales et partielles avec cette structure sont aussi presentes.