Yakun Ju

Ocean University of China, Nanyang Technological University

Papers

Total Citations

H-Index

About

Yakun Ju is a leading researcher in computational imaging and 3D reconstruction, with a focus on photometric stereo and depth estimation. Their work addresses critical challenges in recovering high-precision surface normals and dense depth maps, particularly for complex environments like underwater scenes. Ju’s most cited paper, “A dual-cue network for multispectral photometric stereo” (2019, 27 citations), introduced a novel deep learning framework that leverages multiple spectral cues to improve surface normal reconstruction under varying lighting conditions. Building on this, their 2024 work, “Underwater Surface Normal Reconstruction via Cross-Grained Photometric Stereo Transformer” (23 citations), pioneers the use of transformer architectures for high-resolution 3D data acquisition in underwater robotics, enabling accurate mapping of textureless objects such as seabeds and pipelines. Ju also contributed to self-supervised depth completion with an attention-based loss function (2020), enhancing dense depth prediction for autonomous driving and virtual reality. With a growing citation impact and a focus on real-world applications—from underwater exploration to robotics—Ju’s research is shaping the future of 3D vision in challenging environments.