Adaptive Ultrasound Tissue Harmonic Imaging Based on an Improved Ensemble Empirical Mode Decomposition Algorithm.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Signal Processing

Background:

• Accurate separation of harmonic components in ultrasonic radio frequency (RF) echo signals is crucial for enhancing harmonic imaging quality.
• Existing methods like S_HPF (subjective high-pass filtering) and S_PI (pulse inversion) have limitations, including subjectivity and motion artifacts, respectively.

Purpose of the Study:

• To introduce a novel adaptive method, S_CEEMDAN (complete ensemble empirical mode decomposition with adaptive noise), for separating second harmonic components in ultrasound tissue harmonic imaging.
• To evaluate the performance of S_CEEMDAN against established methods in both simulations and dynamic biological measurements.

Main Methods:

• The S_CEEMDAN method adaptively determines the ensemble size for CEEMDAN based on white noise standard deviation.
• Intrinsic mode functions (IMFs) are extracted from RF echo signals using CEEMDAN and further decomposed based on spectral analysis.
• Fundamental and harmonic components are separated by categorizing IMFs, and then reconstructed by signal accumulation.

Main Results:

• Simulation experiments showed S_CEEMDAN results comparable to S_HPF and superior to S_PI.
• For dynamic carotid artery measurements, S_CEEMDAN significantly improved harmonic images, increasing contrast, contrast-to-noise ratio (CNR), and tissue-to-clutter ratio (TCR) compared to S_HPF and S_PI.

Conclusions:

• The S_CEEMDAN method offers improved adaptivity and reduced motion artifacts, leading to enhanced harmonic images in ultrasound tissue harmonic imaging.
• S_CEEMDAN presents a promising alternative to current methods for achieving higher quality ultrasonic harmonic imaging.