James Sha

Papers

Total Citations

339

H-Index

About

James Sha is a leading researcher in computer vision and 3D object modeling, whose work bridges the gap between raw sensor data and high-fidelity digital representations. He is best known as the principal creator of **BigBIRD**, a landmark large-scale 3D object instance database that has amassed over **322 citations**. This dataset revolutionized the field by providing not just 2D images, but also precise 3D geometry and pose information—a critical resource that enabled breakthroughs in instance recognition, robotic grasping, and manipulation. In complementary work, Sha developed a novel fusion technique that combines range sensor data with silhouette information to produce high-quality 3D scans from commodity RGB-D cameras. This approach directly addressed the limitations of modern reconstruction methods, making professional-grade 3D modeling accessible for applications spanning virtual reality to e-commerce. His contributions have fundamentally shaped how researchers and engineers approach object understanding, providing both the foundational datasets and the algorithmic innovations that power today’s most advanced perception systems.