CIALAB members Dr. Kamel Belkacem-Boussaid and Jeff Prescott have had their recent work accepted for presentation at SPIE 2010
CIALAB members Dr. Kamel Belkacem-Boussaid and Jeff Prescott have had their recent work accepted for presentation at SPIE 2010, taking place February 13-18 in San Diego, California. The accepted papers are:
1- K. Belkacem-Boussaid, J. Prescott, G. Lozanski, and M. Gurcan, “ Segmentation of follicular regions on H&E slides using a matching filter and active contour model”
ABSTRACT: Follicular Lymphoma (FL) accounts for 20-25% of non-Hodgkin lymphomas in the US. The first step in follicular lymphoma grading is the identification of follicles. The goal of this paper is to develop a technique to segment follicular regions in H&E stained images. The method is based on a robust active contour model in which the centroid of the expanded curve is selected manually by the user. The novel aspect of this method is the introduction of matched filter for the flattening of background in the L channel of the Lab color space. The performance of the algorithm was tested by comparing it against the manual segmentations of trained readers using the Zijbendos similarity index. The mean accuracy of the final segmentation compared to the manual ground truth is 0.71 and with a standard deviation of 0.12.
2- J. Prescott, F. Haq, T.M. Best, R. Jackson, and M. Gurcan, “An analysis of methods for the selection of atlases for use in medical image segmentation”
ABSTRACT: The use of atlases has been shown to be a robust method for segmentation of medical images. In this paper we explore different methods of selection of atlases for the segmentation of the quadriceps muscles in MR images, although the results are pertinent for a wide range of applications. First, a set of readers were assigned the task of selecting atlases from a training population of images which were felt to be representative subgroups of the total population. This task was performed with no external knowledge besides the end goal of segmentation. Second, the same readers were given a subset of the training population stratified into modes of the population from which to select templates. Finally, every image in the training set was employed as an atlas, with no input from the readers, and the atlas which had the best initial registration, judged by an appropriate registration metric, was used in the final segmentation procedure. The results show that, for four out of five readers, the inclusion of modal information into the atlas selection process improved the final segmentation. However, the use of every image in the training set as an atlas far outperformed the manual atlas selection method, whether with or without modal information.
The CIALAB members would like to thank all their collaborators for their help in producing these papers.
1- K. Belkacem-Boussaid, J. Prescott, G. Lozanski, and M. Gurcan, “ Segmentation of follicular regions on H&E slides using a matching filter and active contour model”
ABSTRACT: Follicular Lymphoma (FL) accounts for 20-25% of non-Hodgkin lymphomas in the US. The first step in follicular lymphoma grading is the identification of follicles. The goal of this paper is to develop a technique to segment follicular regions in H&E stained images. The method is based on a robust active contour model in which the centroid of the expanded curve is selected manually by the user. The novel aspect of this method is the introduction of matched filter for the flattening of background in the L channel of the Lab color space. The performance of the algorithm was tested by comparing it against the manual segmentations of trained readers using the Zijbendos similarity index. The mean accuracy of the final segmentation compared to the manual ground truth is 0.71 and with a standard deviation of 0.12.
2- J. Prescott, F. Haq, T.M. Best, R. Jackson, and M. Gurcan, “An analysis of methods for the selection of atlases for use in medical image segmentation”
ABSTRACT: The use of atlases has been shown to be a robust method for segmentation of medical images. In this paper we explore different methods of selection of atlases for the segmentation of the quadriceps muscles in MR images, although the results are pertinent for a wide range of applications. First, a set of readers were assigned the task of selecting atlases from a training population of images which were felt to be representative subgroups of the total population. This task was performed with no external knowledge besides the end goal of segmentation. Second, the same readers were given a subset of the training population stratified into modes of the population from which to select templates. Finally, every image in the training set was employed as an atlas, with no input from the readers, and the atlas which had the best initial registration, judged by an appropriate registration metric, was used in the final segmentation procedure. The results show that, for four out of five readers, the inclusion of modal information into the atlas selection process improved the final segmentation. However, the use of every image in the training set as an atlas far outperformed the manual atlas selection method, whether with or without modal information.
The CIALAB members would like to thank all their collaborators for their help in producing these papers.
Comments
Post a Comment