Mojtaba Zarei

As a member of the Center of Virtual Imaging Trials (CVIT), my goal is to enhance the quality of medical images through virtual acquisition. By utilizing virtual patients and scanners, we can gather enough training datasets and target references without jeopardizing patient privacy or well-being. At CVIT, I concentrate on refining the protocol selection guidelines to improve radiologist image evaluation. I have also created physics-informed Deep Neural Networks (DNNs) to minimize image variability caused by scanning processes from various centers/systems. Additionally, I've developed automated pipelines that extract relevant biomarkers to enable precise and reliable quantification of medical images.

I focus on the importance of safety-critical systems, such as medical devices or autonomous vehicles, as they directly impact human well-being. With the ever-evolving technologies embedded in these systems, each component plays a crucial role in ensuring the final outcome is safe and reliable. Through my research, I specialize in creating stochastic model-checking algorithms specifically designed for learning-enabled systems like robots, vehicles, and power systems. These algorithms assess whether a complex system can meet the desired specifications and rely on a unique stochastic model checking approach to gauge the system's reliability within a given confidence interval.

Controlling autonomous vehicles requires a combination of human and intelligence units. Thus, such systems face a number of challenges, including controlling the mobile device smoothly, avoiding obstacles in dynamic environments, and combining sensory data so that the most useful and informative information is used to make a proper decision. For my master's thesis, I developed several obstacle avoidance algorithms and developed a design criteria to synthesize the anti-oscillation controllers. Additionally, designing real-world robots and working with real commercial robots helps me to develop data fusion algorithms that preprocess sensory data and use them for more accurate system feedback.