Exploring impacts of parameters on Ultrasound Localization Microscopy using simulated datasets

Ultrasound imaging is useful for diagnosing various diseases since it provides a real-time view of blood vessels and tissues. Because of its non-invasiveness and capacity to record dynamic physiological processes, it is an essential tool for clinical research. The use of contrast agents has improved its ability to perform in-depth vascular assessments and given researchers a better understanding of microvascular architecture. However, the diffraction barrier, which limits spatial resolution, limits conventional ultrasound imaging. Ultrasound Localization Microscopy (ULM) has emerged as an alternative to these restrictions by striking an equilibrium between spatial resolution and diffraction limit. However, addressing the trade-offs between noise (clutter noise), temporal, and spatial resolution makes ULM implementation difficult. This paper aims to learn how the Number of Particles per Frame, Minimum Length, Maximum Gap Closing, and Maximum Linking Distance affect the resolution, accuracy, and simulated noise levels in the final ULM image by methodically varying them. This study is going to provide insights on fine-tuning ULM for better imaging quality, potentially leading to more accurate and detailed visualization of microvascular structures in clinical settings. Our accuracy has been improved to 64% in this work, up from 55% originally. Furthermore, the objectives of increasing precision, decreasing noise, and improving overall efficiency have been effectively accomplished, resulting in a better comprehension of the microvascular structure.