Papers
8
Total Citations
580
H-Index
6
About
Xin Fan is a prolific researcher whose work sits at the intersection of computer vision, robotic perception, and multi-modal learning, with particular expertise in underwater machine vision and autonomous systems. Fan has made landmark contributions to underwater robotics by developing pioneering datasets and detection frameworks, most notably the UDD dataset — the first 4K HD underwater open-sea farm collection — which has become a foundational resource for training robots to identify and grab marine species such as sea cucumbers, urchins, and scallops. His work on Poisson GAN and AquaNet (109 citations) further demonstrates his drive to bridge data scarcity and real-world deployment in open-sea environments. Fan has also tackled the notoriously difficult problem of underwater depth estimation and color correction through unsupervised adaptation networks (104 citations), advancing the reliability of robotic perception in degraded visibility conditions. More recently, his research has expanded into multi-modality image fusion and segmentation for autonomous driving, with a 2023 paper accumulating an impressive 235 citations, reflecting immediate and widespread community impact. Across his career, Fan's work exemplifies a commitment to building both the datasets and the deep learning architectures that push robotic vision closer to real-world viability.
Research Focus
Key Achievements
Top Papers
- 1
- 2A New Dataset, Poisson GAN and AquaNet for Underwater Object Grabbing109 citations · 2021
- 3
- 4Underwater Species Detection using Channel Sharpening Attention74 citations · 2021
- 5
- 6
- 7A Dataset and Benchmark of Underwater Object Detection for Robot Picking6 citations · 2021
- 8