Xiangyuan Lan

Hong Kong Baptist University, Peng Cheng Laboratory

Papers

Total Citations

H-Index

About

Xiangyuan Lan is a leading researcher in computer vision and deep learning, with a primary focus on facial expression recognition and visual place recognition. His innovative work bridges the gap between computational efficiency and real-world applicability. Lan’s most cited paper, “Convolution by Multiplication: Accelerated Two-Stream Fourier Domain Convolutional Neural Network for Facial Expression Recognition” (2021, 27 citations), introduces a groundbreaking method that leverages Fourier domain operations to dramatically speed up CNNs, achieving state-of-the-art performance in interpreting nonverbal human communication—a critical advance for psychology, human-computer interaction, and robotics. More recently, his 2024 study “Deep Homography Estimation for Visual Place Recognition” (17 citations) tackles the fundamental challenge of robot localization and augmented reality by enhancing hierarchical VPR methods, which balance accuracy and efficiency through refined global feature retrieval. Lan’s contributions are notable for their practical impact, offering computationally efficient solutions that push the boundaries of how machines perceive and navigate the world. His work continues to inspire students and researchers seeking to optimize deep learning architectures for real-time, high-stakes applications.