Jonathan Bonilla

I am a graduate student in Computer Science focused on advancing medical image segmentation. I am building a robust SAM2 model tailored for medical applications, and training it on a diverse range of medical imaging datasets. My work leverages the SAM2 framework, modified by integrating a backbone that combines CNN and VRWKV blocks to capture both local and global features. I also implement an uncertainty-guided automated prompting system to refine segmentation accuracy. Through my research, I aim to develop a model that delivers efficient and reliable segmentation performance for real-world healthcare challenges