Reconstruct the Photoacoustic Image Based On Deep Learning with Multi-frequency Ring-shape Transducer Array.

Area of Science:

• Biomedical Imaging
• Optical Imaging
• Ultrasound Imaging

Background:

• Photoacoustic tomography (PAT) integrates optical and ultrasound imaging for enhanced contrast and spatial resolution.
• Traditional reconstruction algorithms like delay-and-sum are standard but have limitations.
• Deep neural networks show promise for improving photoacoustic image reconstruction.

Purpose of the Study:

• To propose a novel neural network framework for photoacoustic imaging reconstruction.
• To evaluate the performance of the framework using multi-frequency ultrasound sensor data.
• To compare the proposed method against conventional reconstruction algorithms.

Main Methods:

• Developed an end-to-end deep neural network for photoacoustic image reconstruction.
• Utilized multi-frequency ultrasound sensor data with varying center frequencies.
• Trained and tested the network using simulated photoacoustic data from segmented vessels.

Main Results:

• The proposed neural network framework demonstrated superior performance over conventional methods.
• Significantly reduced reconstruction time compared to existing algorithms.
• Achieved better image quality and accuracy in numerical simulations.

Conclusions:

• The novel deep neural network framework offers a powerful approach for photoacoustic imaging reconstruction.
• The method's efficiency makes it suitable for real-time imaging applications.
• Multi-frequency data integration enhances the capabilities of neural network-based PAT.