Enxu Li

Huawei Technologies (Canada), University of Toronto

Papers

Total Citations

286

H-Index

About

Enxu Li is a researcher working at the intersection of computer vision, autonomous systems, and robotics, with a particular focus on perception and sensing technologies. Li's most prominent contribution is the development of (AF)²-S3Net, an innovative deep learning architecture for sparse semantic segmentation that combines attentive feature fusion with adaptive feature selection. This work, published in 2021, has garnered over 240 citations, reflecting its significant impact on the autonomous driving and robotics communities, where accurate environmental perception is critical for passenger and pedestrian safety. The approach addresses a fundamental challenge in road scene understanding by extracting rich semantic information from sparse sensor data — a key bottleneck in real-world deployment of self-driving systems. Beyond autonomous perception, Li has expanded their research into medical and industrial robotics, contributing to the field of continuum robot shape sensing with MoSS, a monocular shape sensing framework published in 2023. This work tackles the demanding challenge of real-time, accurate shape estimation for flexible continuum robots — devices increasingly valued for minimally invasive medical procedures. With a growing citation record across diverse subfields, Li demonstrates a versatile research profile bridging intelligent perception, machine learning, and advanced robotics applications.