Real-time video-rate perfusion imaging using multi-exposure laser speckle contrast imaging and machine learning

Area of Science:

• Biomedical optics
• Microcirculation research
• Medical imaging technology

Background:

• Multi-exposure laser speckle contrast imaging (MELSCI) offers superior accuracy in estimating microcirculatory blood perfusion compared to single-exposure LSCI.
• Previous MELSCI techniques were limited to offline processing due to high data throughput and complex algorithms.

Purpose of the Study:

• To develop and present a novel MELSCI system capable of continuous data acquisition and processing.
• To enable real-time, high-accuracy video-rate perfusion imaging of microcirculation.

Main Methods:

• Implemented the MELSCI algorithm on a field-programmable gate array (FPGA) interfaced with a high-speed CMOS sensor for real-time computation.
• Utilized an artificial neural network trained on simulated data for real-time estimation of perfusion images.
• Validated the system through quantitative phantom experiments and qualitative in vivo studies, comparing MELSCI to single-exposure metrics.

Main Results:

• The real-time MELSCI system achieves high-quality perfusion imaging at 15.6 frames per second.
• MELSCI perfusion demonstrated enhanced spatial and temporal signal dynamics, resolving heartbeat-related variations with greater detail.
• Results showed MELSCI is less susceptible to noise and exhibits higher linearity with laser Doppler perfusion (R² = 0.992) in phantom studies.

Conclusions:

• The developed MELSCI system successfully enables real-time acquisition and processing of perfusion data.
• This advancement overcomes previous limitations, paving the way for dynamic, high-fidelity microcirculatory monitoring.