Tien Do

Papers

Total Citations

H-Index

About

Tien Do is a computer vision researcher whose work focuses on geometric understanding and robust perception in challenging imaging conditions. His most-cited contribution, "Surface Normal Estimation of Tilted Images via Spatial Rectifier" (2020), has garnered 32 citations and introduces a novel approach to inferring 3D surface orientation from images captured at extreme angles. By developing the spatial rectifier module, Do addresses a critical limitation in traditional depth and normal estimation methods, which often fail under perspective distortion. This work has practical implications for augmented reality, robotics, and autonomous navigation, where cameras frequently operate in non-ideal orientations. Do’s research bridges the gap between 2D image analysis and 3D scene reconstruction, demonstrating how geometric priors can enhance neural network performance. His contributions are particularly valuable for students and researchers exploring the intersection of deep learning and geometric computer vision, offering a principled solution to a long-standing problem in the field.