著者

室 伊三男 神谷 陽 本田 真俊 堀江 朋彦

出版者

公益社団法人 日本放射線技術学会

雑誌

日本放射線技術学会雑誌 (ISSN:03694305)

巻号頁・発行日

vol.63, no.1, pp.91-96, 2007-01-20 (Released:2007-02-27)

参考文献数

7

被引用文献数

6 12

---

Echo planar imaging (EPI) is highly sensitive to static magnetic field inhomogeneities. The degree of local image compression and stretching is a function of the static field gradient in the phase-encoding direction. This is caused by the accumulation of a phase shift. Any static field shift will lead to a position shift in the image, and it is the regions with large static fields that are the most difficult to correct. We reduce image distortion by SENSE with an array coil. However, we often use a surface coil because we cannot use an array coil in clinical studies. In this case, image distortion becomes greater, and reduction of distortion is very important. For the purpose of this study, we examined the relation between imaging parameters and image distortion. Image distortion of EPI is unrelated to the following parameters: number of phase encodings, half scan, echo time, and diffusion b-value. However, the following parameters influenced image distortion: FOV, number of frequency encodings, rectangle FOV, and multi-shot imaging. Image distortion of EPI is decided by the area of the phase-encoding gradient and the interval of readout gradients. We hope that many institutions will find these data useful.

著者

小林 明日香 渋川 周平 高野 晋 室 伊三男

出版者

公益社団法人 日本放射線技術学会

雑誌

日本放射線技術学会雑誌 (ISSN:03694305)

巻号頁・発行日

vol.74, no.10, pp.1180-1185, 2018 (Released:2018-10-20)

参考文献数

12

---

We have found that the number of packages influences contrast for brain tissue signals on fluid-attenuated inversion recovery (FLAIR). The purpose of this study was to evaluate the contrast of white and gray matters by changing the number of packages. In a volunteer study (n=8), FLAIR images were obtained with the various number of packages (number of package=2, 3, 4, 5). We investigated the same imaging condition at both 1.5 and 3.0T. The signal intensity of white and gray matters in all volunteers was increased as increasing the number of packages. Moreover, the contrast ratio between white and gray matters was slightly decreased. In our conclusion, the contrast between the gray and white matters on FLAIR was influenced by the number of packages.

著者

室 伊三男 清水 俊太郎 塚本 ひかり

出版者

公益社団法人 日本放射線技術学会

雑誌

日本放射線技術学会雑誌 (ISSN:03694305)

巻号頁・発行日

vol.78, no.1, pp.13-22, 2022

被引用文献数

3

---

<p>【目的】深層学習によるモーションアーチファクト(以下,アーチファクト)削減のアプローチが脳MR画像に有効かを検証する.【方法】本研究ではアーチファクトを含んだ画像と含んでいない画像が学習データとして大量に必要である.臨床画像で学習データを集めるには多くの労力と時間を要して困難である.われわれは脳のアーチファクト画像をシミュレーションによって作成した.ボランティア20人の動きのない頭部画像を取得し,この画像を使用してアーチファクトの異なる画像をシミュレーションによって作成して深層学習を行う.得られた学習モデルのアーチファクト除去効果の検証は,別途テストデータを作成し,テストデータの入力画像と出力画像間のピーク信号対雑音比(peak signal-to-noise ratio: PSNR)と構造的画像類似性(structural similarity: SSIM)を3種類のデノイジング手法で比較した.使用したニューラルネットワークはU-shaped fully convolutional network(U-Net),denoising convolutional neural network(DnCNN)とwide inference network and 5 layers Residual learning and batch normalization(Win5RB)である.【結果】アーチファクト除去効果はU-Netが最も高く,SSIMの平均値は0.978, PSNRの平均値は32.5であった.【結語】本法は脳MRI画像の画質を劣化させずにアーチファクトを軽減できる有効な方法である.</p>

著者

塚本 ひかり 室 伊三男

出版者

公益社団法人 日本放射線技術学会

雑誌

日本放射線技術学会雑誌 (ISSN:03694305)

巻号頁・発行日

vol.77, no.5, pp.463-470, 2021 (Released:2021-05-20)

参考文献数

8

被引用文献数

2

---

Purpose: We focused on deep learning for a reduction of motion artifacts in MRI. It is difficult to collect a large number of images with and without motion artifacts from clinical images. The purpose of this study was to create motion artifact images in MRI by simulation. Methods: We created motion artifact images by computer simulation. First, 20 different types of vertical pixel-shifted images were created with different shifts, and the amount of pixel shift was set from –10 to 10 pixels. The same method was used to create pixel-shifted images for horizontal shift, diagonal shift, and rotational shift, and a total of 80 types of pixel-shifted images were prepared. These images were Fourier transformed to create 80 types of k-space data. Then, phase encodings in these k-space data were randomly sampled and Fourier transformed to create artifact images. The reproducibility of the simulation images was verified using the deep learning network model of U-net. In this study, the evaluation indices used were the structural similarity index measure (SSIM) and peak signal-to-noise ratio (PSNR). Results: The average SSIM and PSNR for the simulation images were 0.95 and 31.5, respectively; those for the clinical images were 0.96 and 31.1, respectively. Conclusion: Our simulation method enables us to create a large number of artifact images in a short time, equivalent to clinical artifact images.