Jinghan Zhang

Papers

Total Citations

H-Index

About

Jinghan Zhang is a researcher advancing the frontiers of geometric deep learning and spherical image analysis. Their most notable contribution is the development of UVS-CNNs, a novel framework for constructing general convolutional neural networks on quasi-uniform spherical images. This work, published in 2024, addresses a critical challenge in processing spherical data—where traditional CNNs falter due to distortion and irregular sampling—by introducing a method that maintains computational efficiency while preserving spatial relationships across the sphere. Though early in its citation trajectory, UVS-CNNs has already garnered 4 citations, signaling its potential impact on fields ranging from omnidirectional vision and 3D scene understanding to climate modeling and astrophysics. Zhang’s research sits at the intersection of computer vision, geometry, and machine learning, offering practical tools for applications where planar assumptions break down. By enabling CNNs to operate effectively on non-Euclidean domains, they are helping to unlock new capabilities in autonomous systems, remote sensing, and virtual reality. Their work exemplifies the growing importance of adapting deep learning architectures to the structure of real-world data, making Zhang a promising voice in the evolution of geometric AI.