| 大会名称 |
|---|
| 2021年 情報科学技術フォーラム(FIT) |
| 大会コ－ド |
| F |
| 開催年 |
| 2021 |
| 発行日 |
| 2021-08-12 |
| セッション番号 |
| 4h |
| セッション名 |
| バイオ情報学と医用画像 |
| 講演日 |
| 2021/08/26 |
| 講演場所(会議室等) |
| h |
| 講演番号 |
| G-012 |
| タイトル |
| Attenuation Correction Factors Generation of a Dedicated Brain PET System with Time-of-Flight Information Using a High-resolution Compact Network |
| 著者名 |
| Tuo Yin, Takashi Obi, |
| キーワード |
| positron emission tomography (PET), attenuation correction, convolutional neural network (CNN), deep learning |
| 抄録 |
| Attenuation correction of annihilation photons is essential in PET image reconstruction process for providing accurate quantitative information on the activity maps. In the absence of an aligned CT device to obtain attenuation information in a dedicated brain PET system, we extract attenuation correction factors (ACF) directly from Time-of-Flight (TOF) PET emission data using a High-resolution Compact Network (HC-Net). HC-Net utilizes 2-D small kernels, dilated convolution and modified residual connections to reduce parameters and enlarge the receptive field. HC-Net generates more accurate ACF in comparison to MLACF algorithm with a lower NRMSE and higher SSIM and PSNR. |
| 本文pdf |
| PDF download (647.7KB) |