A real-time scan conversion algorithm on commercially available microprocessors

Area of Science:

• Medical Imaging
• Computer Science
• Signal Processing

Background:

• Ultrasound scan conversion is crucial for image display.
• Existing algorithms face limitations in processing speed and efficiency on modern hardware.
• Efficient scan conversion is necessary for real-time medical imaging applications.

Purpose of the Study:

• To develop a highly efficient ultrasound scan conversion algorithm for modern microprocessors.
• To optimize the utilization of processing units and direct memory access (DMA) controllers.
• To achieve scan conversion at video rates.

Main Methods:

• Developed a novel scan conversion algorithm leveraging concurrent address calculation, I/O data loading, and interpolation.
• Utilized a direct memory access (DMA) controller for independent handling of I/O operations.
• Exploited superscalar and VLIW processor capabilities for parallel operations.
• Implemented the algorithm on the Texas Instruments TMS320C80 Multimedia Video Processor (MVP).

Main Results:

• The algorithm efficiently dedicates processing power to pixel interpolations.
• It avoids non-productive cycles for address calculations and blanking.
• Achieved an execution time of 11.4 ms for 101,829 nonzero output pixels.
• Demonstrated substantial improvement over previous scan conversion methods.

Conclusions:

• The new algorithm offers significant efficiency gains for ultrasound scan conversion.
• Optimized implementation enables modern processors to support scan conversion at video rates.
• This advancement has the potential to enhance real-time ultrasound imaging systems.