Conference Paper

Multi-mode Narrow-band Thresholding with Application in Liver Segmentation from Low-contrast CT Images

Coll. of Inf. Sci. & Eng., Ritsumeikan Univ., Kusatsu, Japan
DOI: 10.1109/IIH-MSP.2009.78 Conference: Intelligent Information Hiding and Multimedia Signal Processing, 2009. IIH-MSP '09. Fifth International Conference on
Source: IEEE Xplore

ABSTRACT Segmentation of liver in CT images is regarded as a challenge in image processing due to low-contrast of datasets, variety of liver shape, and its non-uniform texture; especially for abnormal cases. In this paper, we deal with normal and abnormal datasets as images containing two or more Gaussian components. We threshold a slice in a narrow band of each mode, find liver pixels based on a priori knowledge, prepare a probability map, and threshold the map to find initial liver border. Final boundary of liver is obtained through a few iterations of `Geodesic Active Contour'. The proposed method was tested on 30 normal and 17 abnormal datasets each containing 159-263 slices; acquired from different CT machines. The results for normal and abnormal datasets are completely acceptable, according to the evaluation done by a specialist. However, for severely abnormal datasets, the proposed method is regarded as a promising algorithm for liver segmentation.

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Recently a growing interest has been seen in minimally invasive treatments with open configuration magnetic resonance (Open-MR) scanners. Because of the lower magnetic field (0.5T), the contrast of Open-MR images is very low. In this paper, we address the problem of liver segmentation from low-contrast Open-MR images. The proposed segmentation method consists of two steps. In the first step, we use K-means clustering and a priori knowledge to find and identify liver and non-liver index pixels, which are used as “object” and “background” seeds, respectively, for graph-cut. In the second step, a graph-cut based method is used to segment the liver from the low-contrast Open MR images. The main contribution of this paper is that the object (liver) and background (non-liver) seeds (regions) in every low-contrast slice of the volume can be obtained automatically by K-means clustering without user interaction.
    Advances in Neural Networks - ISNN 2010, 7th International Symposium on Neural Networks, ISNN 2010, Shanghai, China, June 6-9, 2010, Proceedings, Part II; 01/2010
  • [Show abstract] [Hide abstract]
    ABSTRACT: Successful liver surgery requires an understanding of the patient's particular liver characteristics, including shape and vessel distribution. In clinical medicine, there is a high demand for surgical assistance systems to assess individual patients. Our aims in this study were to segment the liver based on computed tomography volume data and to develop surgical plans for individual patients. The hepatic vessels were semi-automatically extracted from the segmented liver images, and the 3D shape of the liver and extracted vessel distribution were visualized using a surgical simulation system. The 3D visualization of the liver allowed easy recognition of vessel and tumor location and selection of these structures with the 3D pointing device. The surgeon's prior knowledge and clinical experience were integrated into the visualization system to create a practical virtual surgery, leading to improved functionality and accuracy of information recognition in the surgical simulation system. The 3D visualization demonstrated details of individual liver structure, resulting in better understanding and practical surgical simulation.
    Journal of Gastrointestinal Surgery 06/2013; · 2.36 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Successful liver surgery requires a clear understanding of the differences in liver shapes and vessel distribution in different individuals. Furthermore, in clinical medicine, there is a high demand for surgical assistance systems for individual patients. Therefore, we aim to segment the liver on the basis of the CT volume data, semi-automatically extract the vessels from the segmented livers and then visualize the 3D shape and the extracted vessel distribution using a virtual operation system. In addition, to improve the operability and accuracy of information recognition in the virtual operation system, prior knowledge and the clinical experiences of doctors are integrated into the visualization system for a practical virtual surgery. A 3D visualization of the liver, allows the user to easily recognize abnormal regions, which need to be removed, and to simply select this region using a 3D pointing device. Furthermore, 3D visualization, allows details in the structure of the human liver to be better understood and a more practical surgical simulation system can be implemented in our developed system.
    Computer Sciences and Convergence Information Technology (ICCIT), 2011 6th International Conference on; 01/2011

Full-text (2 Sources)

Available from
May 17, 2014