Yunlu Chen's homepage


# About

I am a Ph.D. candidate at University of Amsterdam. My advisor is Dr. Efstratios Gavves. I also received supervision from Dr. Thomas Mensink and Prof. Arnold Smeulders. Since 2020 I have been working closely with Dr. Basura Fernando (A*STAR, Singapore) and Dr. Hakan Bilen (University of Edinburgh). I work on 3D computer vision and deep learning.

I obtained my M.Sc. (cum laude) in artificial intelligence from Utrecht University. My thesis research was conducted at CWI Amsterdam, advised by Dr. Eric Pauwels. I hold a B.Sc. in physics at Fudan University.


# Research Interest

I am interested in deep learning and computer vision in general.

More specifically, my PhD research focuses on 3D deep learning, including recognition or reconstruction of various types of data such as 2.5D RGB-D semantic segmentation and point cloud learning. Currently I am working on 3D reconstruction and 3D generative models with neural implicit 3D representations, in particular how implicit functions can generalise to a shared representation for a variety of shapes. Another direction I am working on is the intersection of graph neural networks and 3D vision.

In addition, I have been also involved in the projects with collaborators and master students on the topics covering transfer learning, self-supervised learning, few-shot learning, image filtering and multi-spectral satellite images.


# Peer-reviewed Publication

[1] Yunlu Chen, Basura Fernando, Hakan Bilen, Thomas Mensink, Efstratios Gavves. "Neural Feature Matching in Implicit 3D Representations. " International Conference on Machine Learning (ICML), 2021.

[2] Zenglin Shi, Yunlu Chen, Efstratios Gavves, Pascal Mettes, Cees G.M. Snoek. "Unsharp Mask Guided Filtering. " IEEE Transactions on Image Processing (TIP), 2021.

[3] Yunlu Chen*, Vincent Tao Hu*, Efstratios Gavves, Thomas Mensink, Pascal Mettes, Pengwan Yang, and Cees G. M. Snoek. "PointMixup: Augmentation for Point Cloud. " European Conference on Computer Vision (ECCV), 2020. (Spotlight presentation, top 5%)

[4] Yunlu Chen, Thomas Mensink, Efstratios Gavves. "3D Neighborhood Convolution: Learning Depth-Aware Features for RGB-D and RGB Semantic Segmentation." International Conference on 3D Vision (3DV), 2019.


# Contact