Acta Optica Sinica, Volume. 44, Issue 7, 0734002(2024)
Source Blur Elimination in Micro-CT Using Self-Attention-Based U-Net
Fig. 1. Network structure diagram. (a) SU-net structure; (b) convolutional modulation module
Fig. 2. Test result images. (a)-(e) and (k)-(o) are ground truth image, FBP, NSM, ESRGAN, and SU-net reconstruction images, respectively; (f)-(j) are partial enlarged images of (a)-(e), respectively; (p)-(t) are partial enlarged images of (k)-(o), respectively. Display window is [0, 0.35] cm-1
Fig. 4. Experimental results outside dataset. (a)-(e) are ground truth image, FBP, NSM, ESRGAN, and SU-net reconstruction images, respectively; (f)-(j) are partial enlarged images of (a)-(e), respectively. Display window is [0, 0.35] cm-1
Fig. 5. Results of actual experiment. (a)-(d) are reconstruction results of FBP, NSM, ESRGAN, and SU-net in line 600, respectively; (e)-(h) are partial enlarged images of (a)-(d), respectively; (i)-(l) are reconstruction results of FBP, NSM, ESRGAN, and SU-net in line 650, respectively; (m)-(p) are partial enlarged images of (i)-(l), respectively. Display window is [0, 0.35] cm-1
|
|
Get Citation
Copy Citation Text
Chuanjiang Liu, Ao Wang, Genyuan Zhang, Wei Yuan, Fenglin Liu. Source Blur Elimination in Micro-CT Using Self-Attention-Based U-Net[J]. Acta Optica Sinica, 2024, 44(7): 0734002
Category: X-Ray Optics
Received: Nov. 29, 2023
Accepted: Jan. 25, 2024
Published Online: Apr. 11, 2024
The Author Email: Liu Fenglin (liufl@cqu.edu.cn)