Hybrid-Supervised Deep Learning for Domain Transfer 3D Protoacoustic Image Reconstruction
Arxiv
|August 23, 2023
Summary
A new deep learning method improves protoacoustic imaging for proton therapy dose verification by addressing limited view artifacts. This enhances accuracy and speed for real-time 3D dose verification in prostate cancer treatment.
Area of Science:
- Medical Imaging
- Physics
- Artificial Intelligence
Background:
- Protoacoustic imaging offers real-time 3D dose verification for proton therapy.
- Limited acquisition angles in protoacoustic imaging cause artifacts, reducing accuracy.
Conclusions:
- The developed deep learning method effectively addresses limited view issues in protoacoustic imaging.
- The approach significantly enhances accuracy and detail in 3D dose verification for proton therapy.
- Feasible for online 3D dose verification in prostate proton therapy due to speed and accuracy.
