Convolutional neural network-based approach to estimate bulk optical properties in diffuse optical tomography.

Area of Science:

• Biomedical optics
• Medical imaging
• Computational imaging

Background:

• Accurate estimation of bulk optical properties is crucial for image quality in diffuse optical tomography (DOT).
• Biological tissues are highly scattering media, posing challenges for optical property estimation.

Purpose of the Study:

• To develop and validate a novel deep learning approach for estimating bulk optical properties in diffuse optical tomography.
• To assess the performance of the proposed method against existing techniques.

Main Methods:

• A convolutional neural network (CNN)-based deep learning model was developed.
• The method was validated using both experimental and simulated diffuse optical tomography data.
• Performance was evaluated by comparing with existing optical property estimation approaches.

Main Results:

• The proposed CNN-based method demonstrated superior estimation accuracy for bulk optical properties compared to existing methods.
• The novel approach achieved a lower computation time.
• Validation with experimental and simulated data confirmed the method's effectiveness.

Conclusions:

• Deep learning, specifically CNNs, offers a powerful and efficient tool for bulk optical property estimation in diffuse optical tomography.
• The developed CNN approach outperforms traditional methods, paving the way for improved DOT imaging of biological tissues.