Zhe Su
4 Papers
Zhe Su is an academic researcher. The author has contributed to research in topics: Deep learning & Correlation. The author has co-authored 1 publications.
Chat about Author
Papers
MVCT image enhancement using reference-based encoder–decoder convolutional neural network
Shuang Jin,Xiaotong Xu,Zhe Su,Long Tang,Mengxun Zheng,Peiwen Liang,Hua Zhang +6 more
TL;DR: This study proposes a reference-based encoder-decoder CNN (RefED-CNN) to enhance noisy MVCT images using KVCT images as auxiliary references, achieving superior denoising and structural detail preservation compared to other methods on phantom and patient data.
2
Chest tomosynthesis image enhancement by bone suppression using convolutional neural networks with synthetic data
Xiaotong Xu,Qian-Qian Li,Shuang Jin,Zhe Su,Yu Zhang +4 more
TL;DR: A deep learning-based bone suppression model using convolutional neural networks effectively removes occluding bony structures in chest tomosynthesis images, achieving a bone suppression rate exceeding 90% and preserving detailed structures with high SSIM values.
Anatomical features driven dual-attention 3DU-Net for dose distribution prediction of breast cancer
Zhe Su,Xiaotong Xu,Shuang Jin,Mengxun Zheng,Long Tang,Peiwen Liang,Hua Zhang +6 more
TL;DR: This study proposes AF-DA3DU-Net, a deep learning model that predicts 3D dose distribution for breast cancer IMRT by incorporating anatomical features and dual attention mechanism, achieving improved accuracy and efficiency in radiation therapy planning.
LGEANet: LSTM-Global temporal convolution-external attention network for respiratory motion prediction.
Kunpeng Zhang,Jiahong Yu,Jian Liu,Qian Li,Shuang Jin,Zhe Su,Xiaotong Xu,Zhenhui Dai,Xuetao Wang,Hua Zhang +9 more
TL;DR: Wang et al. as discussed by the authors proposed a deep learning framework, named as LSTM-Global Temporal Convolution-External Attention Network (LGEANet), for accurate respiratory tumor motion prediction.