Ultrasound tissue characterization involves investigating the interaction of ultrasound waves with biological tissues by analyzing ultrasound backscattered radiofrequency signals (quantitative ultrasound) and ultrasound images. In recent years, there has been a trend for enhancing ultrasound tissue characterization with artificial intelligence (AI), including machine learning and particularly deep learning techniques. These advanced data-driven models can assist in extracting “intelligently” more information from ultrasound backscattered radiofrequency signals and images. This Research Topic focuses on recent advances in AI-empowered ultrasound tissue characterization methods and techniques for improving disease diagnosis, intervention guidance, and therapy monitoring. Incorporating AI into biomedical ultrasound will represent a new direction for ultrasound tissue characterization, which is a growing interest.
A major issue of conventional ultrasound tissue characterization methods without AI is about the manner they extract information from ultrasound backscattered radiofrequency signals and images. Such a manner can be either single-purpose or single-level. For instance, B-mode imaging extracts the log-compressed envelope amplitude information exclusively, while ultrasound attenuation, elastography, or Nakagami imaging extract the information of acoustic attenuation, tissue elasticity (or time delay), Nakagami parameter, respectively, using a specific mathematical or physical model. In contrast, AI-empowered ultrasound tissue characterization methods use data-driven models to learn more abundant and multi-level information from ultrasound backscattered radiofrequency signals and images. Such learnt information can help improve object detection, disease diagnosis, intervention guidance, and therapy monitoring. For instance, the deep learning models trained on ultrasound backscattered signals have shown great potential in clinical fatty liver characterization.
This Research Topic aims to introduce state-of-the-art methods and techniques on AI-empowered ultrasound tissue characterization. The research topics of interest include:
• AI-empowered quantitative ultrasound;
• AI of biomedical ultrasound images;
• Machine learning with ultrasound backscattered radiofrequency signals;
• Deep learning with ultrasound backscattered radiofrequency signals;
• Machine learning with ultrasound images;
• Deep learning with ultrasound images;
• AI-empowered ultrasound imaging for object detection;
• AI-empowered ultrasound imaging for disease diagnosis;
• AI-empowered ultrasound imaging for intervention guidance;
• AI-empowered ultrasound imaging for therapy monitoring;
• AI-based biomedical ultrasound signal processing;
• AI-based biomedical ultrasound image processing.
We welcome article types including Original Research, Review, Brief Research Reports, and Mini-Review.
Ultrasound tissue characterization involves investigating the interaction of ultrasound waves with biological tissues by analyzing ultrasound backscattered radiofrequency signals (quantitative ultrasound) and ultrasound images. In recent years, there has been a trend for enhancing ultrasound tissue characterization with artificial intelligence (AI), including machine learning and particularly deep learning techniques. These advanced data-driven models can assist in extracting “intelligently” more information from ultrasound backscattered radiofrequency signals and images. This Research Topic focuses on recent advances in AI-empowered ultrasound tissue characterization methods and techniques for improving disease diagnosis, intervention guidance, and therapy monitoring. Incorporating AI into biomedical ultrasound will represent a new direction for ultrasound tissue characterization, which is a growing interest.
A major issue of conventional ultrasound tissue characterization methods without AI is about the manner they extract information from ultrasound backscattered radiofrequency signals and images. Such a manner can be either single-purpose or single-level. For instance, B-mode imaging extracts the log-compressed envelope amplitude information exclusively, while ultrasound attenuation, elastography, or Nakagami imaging extract the information of acoustic attenuation, tissue elasticity (or time delay), Nakagami parameter, respectively, using a specific mathematical or physical model. In contrast, AI-empowered ultrasound tissue characterization methods use data-driven models to learn more abundant and multi-level information from ultrasound backscattered radiofrequency signals and images. Such learnt information can help improve object detection, disease diagnosis, intervention guidance, and therapy monitoring. For instance, the deep learning models trained on ultrasound backscattered signals have shown great potential in clinical fatty liver characterization.
This Research Topic aims to introduce state-of-the-art methods and techniques on AI-empowered ultrasound tissue characterization. The research topics of interest include:
• AI-empowered quantitative ultrasound;
• AI of biomedical ultrasound images;
• Machine learning with ultrasound backscattered radiofrequency signals;
• Deep learning with ultrasound backscattered radiofrequency signals;
• Machine learning with ultrasound images;
• Deep learning with ultrasound images;
• AI-empowered ultrasound imaging for object detection;
• AI-empowered ultrasound imaging for disease diagnosis;
• AI-empowered ultrasound imaging for intervention guidance;
• AI-empowered ultrasound imaging for therapy monitoring;
• AI-based biomedical ultrasound signal processing;
• AI-based biomedical ultrasound image processing.
We welcome article types including Original Research, Review, Brief Research Reports, and Mini-Review.