Disentangling the Frequency Content in Optoacoustics

Area of Science:

• Biomedical Imaging
• Optoacoustic Tomography
• Signal Processing

Background:

• Optoacoustic tomography (OAT) signals contain broadband frequencies encoding multi-scale structural information.
• Concurrent processing of these signals can lead to poor image contrast and fidelity due to low-frequency bias.
• Standard filtering methods are incompatible with entangled frequencies from different structure sizes, causing artifacts.

Purpose of the Study:

• To introduce a novel reconstruction method, frequency-band model-based (fbMB) reconstruction, for high-fidelity OAT imaging.
• To address the challenge of concurrent processing of broadband signals in OAT.
• To improve the rendering of anatomical structures across all physical scales.

Main Methods:

• Developed frequency-band model-based (fbMB) reconstruction to separate frequency-band-specific image components during image formation.
• Utilized soft priors to disentangle overlapping frequency content and optimize localization and structural integrity.
• Applied fbMB to reconstruct optoacoustic signals from structures of varying sizes.

Main Results:

• fbMB reconstruction significantly improved image contrast and fidelity in optoacoustic images.
• Demonstrated unprecedented detail in revealing anatomical structures in in vivo mouse images.
• Showcased enhanced accuracy of spectral unmixing in small vasculature using fbMB-reconstructed images.

Conclusions:

• fbMB reconstruction precisely treats frequency components of optoacoustic signals, improving image quality, accuracy, and quantification.
• This method enables high-fidelity rendering of structures across all scales, overcoming limitations of traditional approaches.
• fbMB offers a superior method for optoacoustic reconstructions, advancing biomedical imaging capabilities.