Varun Gulshan

Papers

Total Citations

412

H-Index

About

Varun Gulshan is a leading researcher in computer vision and medical imaging, best known for his pioneering work in interactive image segmentation and deep learning for healthcare. His most-cited paper, "Geodesic Star Convexity for Interactive Image Segmentation" (2010, 412 citations), introduced a powerful shape constraint that extends star-convexity priors by using geodesic paths and multiple stars, enabling globally optimal segmentation with minimal user input. This work has become a foundational technique in interactive segmentation, influencing both academic research and practical tools. Gulshan’s impact is further underscored by his contributions at Google, where he led the development of deep learning models for diabetic retinopathy screening, achieving clinical-grade accuracy and demonstrating the potential of AI in ophthalmology. His research bridges theoretical advances in optimization and user interaction with real-world medical applications, earning him recognition as a key figure in applied computer vision. With over 400 citations on his seminal segmentation paper alone, Gulshan’s work continues to shape how machines understand and delineate objects in images, particularly in high-stakes healthcare settings.