| Japanese |
|---|
| Title | 核医学画像の不均一さの定量化に関する一試案 |
|---|
| Subtitle | 原著 |
|---|
| Authors | 手島建夫*, 井沢豊春*, 平野富男*, 蝦名昭男*, 白石晃一郎*, 今野淳* |
|---|
| Authors(kana) | |
|---|
| Organization | *東北大学抗酸菌病研究所内科 |
|---|
| Journal | 核医学 |
|---|
| Volume | 22 |
|---|
| Number | 9 |
|---|
| Page | 1367-1376 |
|---|
| Year/Month | 1985/9 |
|---|
| Article | 原著 |
|---|
| Publisher | 日本核医学会 |
|---|
| Abstract | 「要旨」 核医学画像診断においてアイソトープ (RI) 分布の不均一さという表現の中に重要な生理学的あるいは機能的な情報を含んでいると考えられる. しかし, この表現はこれまで判然としないまま定性的な観点でしか理解されてきていないように思われる. この問題への一つのアプローチとして不均一さを数値化し定量的に表現する方法を考案したので, 報告した. すなわちマトリックスデータとして与えられた画像のプロフィールをもとに, 単純な手法で不均一さを表示する指標を定義し, コンピューターにより画像の解析を行った. 水平方向のプロフィールを解析して得られる変数をH変数 (H variables) , 垂直方向をV変数 (V variables) と定義し, 両変数をあわせた場合を, H+V解析 (H+V analysis) として総指標数33個を定義した. 他方, H変数, V変数を区別せず平均的に扱った場合をHV解析 (HV analysis) とし, この時, 21個の指標を定義した. 本法をRI画像に適用した結果, 不均一さを論じるには一画素当たり250カウント以上の計測が必要であるなどの基礎的知見が得られた. |
|---|
| Practice | 臨床医学：一般 |
|---|
| Keywords | Radionuclide image, Unevenness on images, Computer analysis |
|---|
| English |
|---|
| Title | Numerization of Unevenness on Radionuclide Images |
|---|
| Subtitle | Original Articles |
|---|
| Authors | Takeo TESHIMA, Toyoharu ISAWA, Tomio HIRANO, Akio EBINA, Koichiro SHIRAISHI, Kiyoshi KONNO |
|---|
| Authors(kana) | |
|---|
| Organization | Department of Medicine, The Research Institute for Chest Diseases and Cancer, Tohoku University |
|---|
| Journal | The Japanese Journal of nuclear medicine |
|---|
| Volume | 22 |
|---|
| Number | 9 |
|---|
| Page | 1367-1376 |
|---|
| Year/Month | 1985/9 |
|---|
| Article | Original article |
|---|
| Publisher | THE JAPANESE SOCIETY OF NUCLEAR MEDICINE |
|---|
| Abstract | [Summary] The term unevenness or inhomogeneity has been used rather vaguely in nuclear medicine to characterize the radioactive distribution in organ images. The purpose of this study was to establish methodology to numerize the unevenness on images and to find the physiological and functional aspects included in uneven images. Using the count profiles of each of the rows and/or columns derived from image matrix data, many indexes, such as size, deposition characteristics and indexes of unevenness of count distribution were defined and computed. The values derived by analyzing count profiles of each row were named horizontal variables. Similar procedures were performed vertically for each column and the values thus obtained were named vertical variables. To evaluate the unevenness of images, two different approaches were tried. One was H+V analysis : the indexes were taken from those size and deposition characteristics, horizontal and vertical variables and their respective means, standard deviation, coefficient of variation and lengths. Thus the total number of indexes was 33 in H+V analysis. Another approach was HV analysis : here disregarding the rows and columns, similar calculations were made for the combined rows and columns for the H and V variables. The total number of indexes was 21 in HV analysis. We applied this method to lung perfusion images and found that at least more than 250 counts per cell (pixel) were necessary for the analysis of unevenness. |
|---|
| Practice | Clinical medicine |
|---|
| Keywords | Radionuclide image, Unevenness on images, Computer analysis |
|---|