01-10-2010, 11:51 AM
[attachment=4717]
This article is presented by:
FELICIA S. JONES
(Under the Direction of Hamid Arabnia)
ABSTRACT
The National Library of Medicine’s Visible Human Project is a digital image library containing full color anatomical, CT and MR images representing an adult male and female. Segmentation of the Visible Human datasets offers many additions to the original goal of a three-dimensional representation of a computer generated anatomical model of the human body. This paper presents an automatic segmentation algorithm called the Medical Image Segmentation Technique, MIST, which is based on a seeded region growing approach. The technique repeatedly extracts anatomical regions of interest from two-dimensional cross section images to create three-dimensional visualizations of these anatomical organs, bones and tissues. Resulting segmentations of this technique are compared with existing segmentation algorithms. This method proves to produce better whole organ and tissue segmentations than existing algorithms.
INTRODUCTION
The most common images used in medicine are the CT and MR images. X-ray computed tomography (CT, or CAT for computer assisted tomography) involves an Xray source and a detector positioned on opposite sides of a patient. The equipment is arranged in such a way that the X-ray beam can be rotated about one axis while the patient is translated parallel to that axis. In this way, X-ray images of each section are digitally recorded from many angles. Subsequently, algorithms derived from a mathematical procedure called tomography are applied to reconstruct a three-dimensional matrix of values representing the X-ray transmission properties in the volume occupied by the specimen . Magnetic-resonance imaging (MRI) is an imaging technique used primarily in medical settings to produce high quality images of the inside of the human body. A MRI is similar to CT, but it does not use X-rays. Instead, a strong, magnetic field is used to affect the orientation of protons, which behave like miniature magnets and tend to align themselves with the external field. Anatomical images are also used in the medical domain. The National Library of Medicine’s (NLM) Visible Human Project ® (VHP) is the creation of complete, anatomically detailed, three-dimensional representations of the normal male and female human bodies. The male cadaver was sectioned at 1-millimeter intervals, the female 2 cadaver at .33-millimeter intervals. The anatomical images in this dataset are full color 24-bit cross section images that show the detail of the human body [1]. The dataset includes digitized photographic images from cross-sectioning, digital images derived from computerized tomography (CT) and digital magnetic resonance (MR) images of human cadavers. The male dataset consists of 1,871 cross sections for each mode, anatomical and CT, while the female dataset consists of 5,189 anatomical images. Both the male and female datasets contain axial images of the entire body in the anatomical and CT mode; the MR mode contains axial images of the head and neck areas and longitudinal sections of the rest of the body obtained at 4mm intervals [1]. To date, researchers from all over the world have used the dataset for medical diagnostic and treatment applications, as well as for educational purposes for students of all ages .Image segmentation and anatomical feature extraction is a widely researched area related to the use of the dataset.
For more information about this article,please follow the link:
http://docs.googleviewer?a=v&q=cache:PlD...6_4_09.pdf+This+paper+presents+an+automatic+segmentation+algorithm+called+the+Medical+Image+Segmentation&hl=en&gl=in&pid=bl&srcid=ADGEESi-3bYKINtxoLy7yYvqoVZ7MaecF6jGcIe3squ_kHTBOryR5TbnTd8cHdWIYiVaJm45yA7ud_DuaOw8jVVtWRB6vB_3Qm2IfN79uquZ0DSN1nxEWqCMtBZOwb6E8z8Vfut88eH5&sig=AHIEtbTE-5WRq3cbFF7mdJnGpFJBZaRz_w
This article is presented by:
FELICIA S. JONES
(Under the Direction of Hamid Arabnia)
MEDICAL IMAGE SEGMENTATION
ABSTRACT
The National Library of Medicine’s Visible Human Project is a digital image library containing full color anatomical, CT and MR images representing an adult male and female. Segmentation of the Visible Human datasets offers many additions to the original goal of a three-dimensional representation of a computer generated anatomical model of the human body. This paper presents an automatic segmentation algorithm called the Medical Image Segmentation Technique, MIST, which is based on a seeded region growing approach. The technique repeatedly extracts anatomical regions of interest from two-dimensional cross section images to create three-dimensional visualizations of these anatomical organs, bones and tissues. Resulting segmentations of this technique are compared with existing segmentation algorithms. This method proves to produce better whole organ and tissue segmentations than existing algorithms.
INTRODUCTION
The most common images used in medicine are the CT and MR images. X-ray computed tomography (CT, or CAT for computer assisted tomography) involves an Xray source and a detector positioned on opposite sides of a patient. The equipment is arranged in such a way that the X-ray beam can be rotated about one axis while the patient is translated parallel to that axis. In this way, X-ray images of each section are digitally recorded from many angles. Subsequently, algorithms derived from a mathematical procedure called tomography are applied to reconstruct a three-dimensional matrix of values representing the X-ray transmission properties in the volume occupied by the specimen . Magnetic-resonance imaging (MRI) is an imaging technique used primarily in medical settings to produce high quality images of the inside of the human body. A MRI is similar to CT, but it does not use X-rays. Instead, a strong, magnetic field is used to affect the orientation of protons, which behave like miniature magnets and tend to align themselves with the external field. Anatomical images are also used in the medical domain. The National Library of Medicine’s (NLM) Visible Human Project ® (VHP) is the creation of complete, anatomically detailed, three-dimensional representations of the normal male and female human bodies. The male cadaver was sectioned at 1-millimeter intervals, the female 2 cadaver at .33-millimeter intervals. The anatomical images in this dataset are full color 24-bit cross section images that show the detail of the human body [1]. The dataset includes digitized photographic images from cross-sectioning, digital images derived from computerized tomography (CT) and digital magnetic resonance (MR) images of human cadavers. The male dataset consists of 1,871 cross sections for each mode, anatomical and CT, while the female dataset consists of 5,189 anatomical images. Both the male and female datasets contain axial images of the entire body in the anatomical and CT mode; the MR mode contains axial images of the head and neck areas and longitudinal sections of the rest of the body obtained at 4mm intervals [1]. To date, researchers from all over the world have used the dataset for medical diagnostic and treatment applications, as well as for educational purposes for students of all ages .Image segmentation and anatomical feature extraction is a widely researched area related to the use of the dataset.
For more information about this article,please follow the link:
http://docs.googleviewer?a=v&q=cache:PlD...6_4_09.pdf+This+paper+presents+an+automatic+segmentation+algorithm+called+the+Medical+Image+Segmentation&hl=en&gl=in&pid=bl&srcid=ADGEESi-3bYKINtxoLy7yYvqoVZ7MaecF6jGcIe3squ_kHTBOryR5TbnTd8cHdWIYiVaJm45yA7ud_DuaOw8jVVtWRB6vB_3Qm2IfN79uquZ0DSN1nxEWqCMtBZOwb6E8z8Vfut88eH5&sig=AHIEtbTE-5WRq3cbFF7mdJnGpFJBZaRz_w
