医用画像情報学会雑誌
Online ISSN : 1880-4977
Print ISSN : 0910-1543
ISSN-L : 0910-1543
検索
詳細検索
OR
閲覧
検索
21 巻, 1 号
January
選択された号の論文の30件中1～30を表示しています
すべての抄録を表示する
    • |<
    • <
    • 1
    • >
    • >|
特別寄稿
特別講演資料
寄稿
分野別ミニ解説
企業招待解説
CAD 特集一般論文
  • Turn round法を用いた仮想気管支鏡による気管支鏡検査の支援
    小倉 敏裕, 高津 一朗, 佐藤 之俊, 奥村 栄, 平野 純, 尾本 健一郎, 中川 健
    2004 年 21 巻 1 号 p. 103-108
    発行日: 2004年
    公開日: 2005/10/14
    DOIhttps://doi.org/10.11318/mii.21.103
    ジャーナル フリー
    抄録を表示する抄録を非表示にする
    Virtual bronchoscopic images could be easily produced by multi detector-row CT scanner recently. For decision of a path of virtual bronchoscopy from upper end of trachea to mass shadow, we have to explore into an area adjacent to the mass shadow beyond several bronchial bifurcations. Contrary to ordinary decision of route to the mass shadow, we obtained the opposite route from mass shadow to upper end of trachea. And we reverse the obtained opposite route by “Turn round method”. This reversal route was decided as a path of virtual bronchoscopy. This method provided correct path and could be make considerably faster than conventional methods. As the result of several trial cases, this method provides correct path perfectly. But our method has the twist rotation phenomenon of the virtual bronchoscopy on its axis, and we lost the observation direction on the three dimensional luminal images as disadvantage.
    抄録全体を表示
    PDF形式でダウンロード (1670K)
  • モルフォロジー処理を利用した頭部MR画像における小脳および脳幹部の自動抽出法
    林 則夫, 真田 茂, 鈴木 正行, 松浦 幸広
    2004 年 21 巻 1 号 p. 109-115
    発行日: 2004年
    公開日: 2005/10/14
    DOIhttps://doi.org/10.11318/mii.21.109
    ジャーナル フリー
    抄録を表示する抄録を非表示にする
    MR imaging is an important method for the diagnosis of diseases caused by various cerebral pathologies. Assessment of the volume reduction such as cerebral atrophy, SDAT (Senile Dementia of Alzheimer Type) and OPCA (Olivopontocerebellar atrophy) is very important in clinical practice. However, the assessment of the atrophy used to be performed by manual measurement or visual evaluation. Therefore, in order to diagnose by quantitative assessment, it is desirable to measure the regional volume automatically. In this study, we investigated an automated segmentation method of cerebellum and brainstem on MR images using morphological information. An automated method was consisted of the following three steps: (1) segmentation of the brain on MR images (2) segmentation of the cerebellum and brainstem on the brain images using mathematical morphology (3) correction of errors on the segmented images using 3-D information. The results indicated that the regions obtained by automated segmentation method were visually similar to those by manual method. An average of the rate of correctly recognized regions is over 70%. However, an average of the rate of unrecognized regions is over 10%. If segmentation accuracy is improved moreover, our method may provide the quantitative diagnostic information.
    抄録全体を表示
    PDF形式でダウンロード (667K)
  • 動画対応フラットパネルディテクタ装置を用いた新しい膝関節動態解析法
    森岡 茂晃, 真田 茂, 北岡 克彦, 中村 立一, 松井 武司, 井上 仁司
    2004 年 21 巻 1 号 p. 116-121
    発行日: 2004年
    公開日: 2005/10/14
    DOIhttps://doi.org/10.11318/mii.21.116
    ジャーナル フリー
    抄録を表示する抄録を非表示にする
    Recently, a dynamic flat panel detector (FPD) system is appeared as a new digital X-ray modality, and makes it possible to provide dynamic X-ray images for a screening examination of knee function. Knee joint incongruity and osteoarthritis may be diagnosed as the screening examination. We developed a computer analysis method for knee joint movement with using the dynamic FPD. On lateral view of knee joint X-ray images, a movement of the lateral condyle of femur during flexion and extension movement is important in clinical practice, and a relationship between the movement of the lateral condyle of femur and knee joint angles during the movement is also important. We obtained 30 images of five healthy volunteers at 10 seconds during the knee-bending movement. An outline of the lateral condyle of femur were enhanced by the Sobel operator, which selected edge gradients within a limited range of orientations, and then detected by Binarization and the Template matching technique. The knee joint angles were automatically calculated by techniques such as the Sobel operator, and the Hough transform. It concludes that a possibility of computer analysis for a knee joint movement with the dynamic FPD as a new examination was indicated.
    抄録全体を表示
    PDF形式でダウンロード (669K)
  • 微小石灰化像良悪性鑑別のための人工ニューラルネットワーク法とファジィ推論法のROC比較評価
    李 鎔範, 蔡 篤儀
    2004 年 21 巻 1 号 p. 122-130
    発行日: 2004年
    公開日: 2005/10/14
    DOIhttps://doi.org/10.11318/mii.21.122
    ジャーナル フリー
    抄録を表示する抄録を非表示にする
    The purpose of this study is to compare four CAD-algorithms under the same condition, and to verify the most effective algorithm for the classification of microcalcifications on mammogram. The four CAD-algorithms are BP-ANN, GA-ANN, FL and GA-FL. BP-ANN is a conventional artificial neural network with back-propagation learning. GA-ANN is an improved ANN based on genetic algorithm to determine the weighting coefficients at ANN. FL is a conventional fuzzy logic algorithm using gaussian-distributed membership functions. GA-FL is an improved FL based on genetic algorithm to optimize the membership functions. Comparison results are indicated by ROC curves. The Az values of the ROC curves for BP-ANN, GA-ANN, FL and GA-FL were 0.86, 0.80, 0.89 and 0.95, respectively. When sensitivity of each algorithm was 100%, specificities of BP-ANN, GA-ANN, FL and GA-FL were 69%, 54%, 31% and 77%, respectively. Therefore, these results show the effectiveness of GA-FL algorithm. GA-FL has a potential to become useful CAD algorithm to classify microcalcifications on mammogram.
    抄録全体を表示
    PDF形式でダウンロード (783K)
CAD 特集研究速報
  • 脳形状特徴を用いた正中線検出の試み
    歸山 智治, 児玉 直樹, 金子 智之, 福本 一朗
    2004 年 21 巻 1 号 p. 131-136
    発行日: 2004年
    公開日: 2005/10/14
    DOIhttps://doi.org/10.11318/mii.21.131
    ジャーナル フリー
    抄録を表示する抄録を非表示にする
    In this study, we detected the brain stem and the corpus callosum from MRI. The mid line was determined by connecting them in a straight line. We studied 33 subjects, 19 Alzheimer-type dementia patients and 14 healthy elderly controls. The consistent rate of the brain stem was 80.5±5.9% in Alzheimer-type dementia patients, 81.0±3.5% in healthy elderly controls. Moreover, a radiological technologist estimated the mid line, it was detected in 13 of 19 Alzheimer-type dementia patients, and in 8 of 14 healthy elderly controls. It was possible to detect the mid line of both groups, the Alzheimer-type dementia patients and the healthy elderly controls can be by this method. The accuracy of this method improved and it is thought that it will be use in the to automatic diagnosis of Alzheimer-type dementia by comparing a brain on both sides of brain.
    抄録全体を表示
    PDF形式でダウンロード (792K)
  • 複素自己回帰モデルによる凹領域認識を用いた大腸ポリープ像の自動検出
    久山 慶子, 福岡 大輔, 原 武史, 小倉 敏裕, 藤田 広志
    2004 年 21 巻 1 号 p. 137-142
    発行日: 2004年
    公開日: 2005/10/14
    DOIhttps://doi.org/10.11318/mii.21.137
    ジャーナル フリー
    抄録を表示する抄録を非表示にする
    We have been developing a computer-aided detection (CAD) system for detecting colonic polyps in CT colonography. The CAD scheme consists of the following two main steps: (1) producing cross-sectional images along a centerline of colon region and (2) detecting colonic polyps based on the complex autoregressive model for recognition of concave areas. In this paper, we describe an experimental result in applying complex partial correlation (PARCOR) coefficient to CT colonography and evaluate the CAD performance. As an initial result, the sensitivity was 82% (9/11) with 184 false positives (FPs) per case in 7 clinical cases, which indicates that our proposed method may be effective for polyp detection but the issue of existing FP candidates has to be solved in the next step.
    抄録全体を表示
    PDF形式でダウンロード (651K)
論文
    • |<
    • <
    • 1
    • >
    • >|
feedback
 Top