Deep learning-based image reconstruction and post-processing methods in positron emission tomography for low-dose...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Radiology

Background:

• Positron emission tomography (PET) image quality is vital for diagnostics.
• Deep learning (DL) shows great promise for medical image enhancement.
• A rapidly advancing field integrates DL into PET image processing.

Purpose of the Study:

• To review DL methods for PET image reconstruction and post-processing.
• To cover applications in low-dose imaging and resolution enhancement.
• To discuss clinical challenges and future directions.

Main Methods:

• Review of DL integration in PET image reconstruction (regularization, data-driven mapping).
• Review of DL-based post-processing for low-dose, temporal, and spatial resolution enhancement.
• Analysis of conventional PET image processing techniques.

Main Results:

• DL methods offer significant potential for improving PET image quality.
• Integration into reconstruction and post-processing yields enhanced diagnostic capabilities.
• Challenges remain in clinical implementation and future research.

Conclusions:

• Deep learning is a powerful tool for enhancing PET imaging.
• Further research is needed to overcome clinical application barriers.
• DL promises to advance diagnostic accuracy in PET scans.