Deconvolution of ultrasound imaging

Deconvolution of ultrasound imaging enhances axial resolution of ultrasonic images by applying a joint sparse representation model to deconvolve acoustic pulse broadening from limited-bandwidth transducer signals, enabling improved visualization and measurement of layered tissue structures such as vessel walls.

Key Features:

• Joint Sparse Representation Model: Integrates sparse deconvolution along the axial direction with a sparsity-favoring constraint along the lateral direction to leverage information from neighboring scan lines.
• Enhanced Axial Resolution: Connects nearby pixels across scan lines rather than performing per-scan-line deconvolution, yielding improved axial resolution in ultrasonic images.
• Automatic Measurement Capabilities: Demonstrates automatic carotid intima-media thickness measurement with reported precision of 0.56±0.03 mm compared to 0.60±0.06 mm for manual measurement.

Scientific Applications:

• Vascular imaging: Improves visualization of layered structures in blood vessels to support detailed structural assessment.
• Quantitative tissue measurement: Enables more precise measurements such as carotid intima-media thickness for diagnostic and research use.

Methodology:

Applies a joint sparse representation model performing sparse deconvolution along the axial direction combined with a sparsity-favoring lateral constraint that links neighboring scan lines; validated using simulations and real-world data.