Hyperspectral interventional imaging for enhanced tissue visualization and discrimination combining band selection...

Area of Science:

• Medical imaging
• Biomedical optics
• Surgical technology

Background:

• Hyperspectral imaging (HSI) offers noninvasive tissue characterization.
• HSI is increasingly applied in medical settings for enhanced visualization.
• Surgeons often struggle to identify critical anatomical structures with the naked eye.

Purpose of the Study:

• To develop an easily integrated hyperspectral imaging system for operating rooms.
• To enhance the visualization of difficult-to-detect anatomical tissues.
• To optimize band selection for improved tissue contrast in real-time surgical guidance.

Main Methods:

• A Liquid Crystal Tunable Filter (LCTF)-based HSI system operating from 400-1700 nm was utilized.
• Band selection and transformation methods were compared to identify optimal spectral bands.
• Combined band selection approaches were proposed and evaluated using various criteria.

Main Results:

• Proposed combined band selection methods demonstrated superior performance.
• These methods provided rich spectral information and high tissue separability.
• Significant discrimination between biological tissues was achieved with short computational times.

Conclusions:

• A novel HSI system was developed to improve visualization of critical biological tissues.
• The proposed methods offer a favorable balance between performance metrics.
• The system, particularly in the Short-Wave Infrared (SWIR) band, surpasses naked-eye detection capabilities.