A Review on Low-Dose Emission Tomography Post-Reconstruction Denoising with Neural Network Approaches

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Radiology

Background:

• Low-dose emission tomography (ET) is vital for functional imaging but suffers from noise amplification due to photon counting randomness.
• Reducing patient radiation dose in ET is a significant clinical challenge.

Approach:

• This review surveys post-processing techniques for low-dose ET, focusing on deep neural network (NN) applications.
• It examines current advancements and future prospects of NN-based methods in ET image enhancement.

Key Points:

• Deep learning, particularly NNs, shows promise in denoising and improving resolution in low-dose ET.
• NNs can potentially overcome limitations of traditional post-processing methods.
• The review highlights the growing role of artificial intelligence in medical imaging.

Conclusions:

• Deep neural networks offer a powerful approach to enhance low-dose ET image quality and resolution.
• Further research into NN-based techniques will advance diagnostic capabilities in medical imaging.
• AI-driven solutions are poised to significantly impact the future of low-dose ET.