Robust and efficient ultrasound video coding in noisy channels using H.264.

Area of Science:

• Medical Imaging
• Video Compression
• Signal Processing

Background:

• Carotid ultrasound videos require high diagnostic quality for accurate medical assessment.
• Existing video compression techniques may not adequately preserve diagnostic information in medical videos, especially under noisy transmission conditions.
• Flexible Macroblock Ordering (FMO) is a video coding tool that can partition a picture into slices, but its application in medical imaging with variable quality has not been fully explored.

Purpose of the Study:

• To define and utilize diagnostic Regions of Interest (ROIs) for enhanced compression of carotid ultrasound videos.
• To extend Flexible Macroblock Ordering (FMO) for variable quality slice encoding based on diagnostic importance.
• To improve the resilience and diagnostic performance of compressed medical videos transmitted over error-prone channels.

Main Methods:

• Definition of diagnostic Regions of Interest (ROIs) within carotid ultrasound videos.
• Application of Flexible Macroblock Ordering (FMO) with variable quality slice encoding, prioritizing diagnostically important regions.
• Integration of Redundant Slices (RS) to enhance video resilience against packet loss.
• Evaluation using five carotid ultrasound videos at QCIF and CIF resolutions with packet loss rates up to 30%.

Main Results:

• Encoded videos demonstrated enhanced diagnostic performance in noisy environments, confirmed by subjective clinical ratings and objective PSNR metrics.
• The proposed scheme achieved significant reductions in bandwidth demands.
• The method proved effective even at high packet loss rates (up to 30%).

Conclusions:

• The developed ROI-based FMO encoding scheme with variable quality slices and RS effectively enhances diagnostic performance for carotid ultrasound videos.
• This approach offers a valuable solution for transmitting high-quality medical videos over unreliable networks while reducing bandwidth consumption.
• The findings suggest a promising direction for optimizing medical video compression for clinical applications.