3D ultrasound reconstruction algorithms from analog and digital data

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Image Reconstruction

Background:

• Freehand 3D ultrasound is gaining clinical traction for diagnostics and image-guided interventions.
• Existing 3D ultrasound reconstruction methods utilize voxel-, pixel-, or function-based algorithms.
• Historically, digitized analog video has been the primary input for 3D ultrasound reconstruction.

Purpose of the Study:

• To compare the image quality of 3D ultrasound volumes reconstructed from different data sources (processed digital video, unprocessed digital scan-line data, analog video).
• To evaluate the impact of various reconstruction algorithms on 3D ultrasound image quality.
• To determine which data source provides the best reconstruction quality for clinical applications.

Main Methods:

• Utilized the Sonix RP ultrasound scanner and L14-5/38 linear probe for data acquisition.
• Compared reconstruction algorithms and data sources using criteria: wire target detectability, spatial resolution, fine structure detectability, subjective tissue quality, and volume geometry.
• Employed traditional methods by assessing reconstructed volumes after removing a percentage of input data.

Main Results:

• Processed digital video resulted in better reconstruction quality than digital scan-line data.
• Digital video (both processed and unprocessed) outperformed analog video.
• The selection of the video data source had a greater influence on image quality than the choice of reconstruction algorithm.

Conclusions:

• Processed digital video is the preferred data source for high-quality 3D ultrasound reconstruction.
• Optimizing data source selection is crucial for enhancing diagnostic accuracy and intervention guidance in 3D ultrasound.
• Future research should focus on leveraging digital data streams for improved 3D ultrasound performance.