Tian Yu (田煜)

About me

I am currently a master's student in Software Engineering at the Northwest University, advised by Professor (Cao Xin).

I received the B.S. degree from the Northwest University of Software Engineering, in 2020.

Research

My research interests include:

• 3D Point Cloud Analysis: point cloud feature extraction based on deep learning methods.

• Unsupervised Representation Learning: exploring unsupervised (self-supervised) representation learning methods for better transferability on ended downstream tasks.

• Point Cloud Reconstruction: recovering complete point cloud data from vectors.

Current work

• Point-Voxel Network for Point Cloud Analysis

• Masked Aotu-Encoder for Point CLoud Unsupervised Representation Learning

• Weakly-supervised Learning for Point Cloud Segmentation

Publications

1. J. Liu*, Yu Tian*, G. Geng*, H. Wang, D. Song, K. Li, M. Zhou, and X. Cao, "Uma-net: an unsupervised representation learning network for 3d point cloud classification," J. Opt. Soc. Am. A, vol. 39, no. 6, pp. 1085–1094, Jun 2022, (*equal contribution). [Online]. Available: https://opg.optica.org/josaa/abstract.cfm?URI=josaa-39-6-1085 (IF = 2.104) [pdf]

2. J. Liu, G. Geng, Yu Tian, Y. Wang, Y. Liu, and M. Zhou, “Unsupervised representation learning for cultural relics based on local-global bidirectional reasoning,” Optics and Precision Engineering, vol. 30, no. 18, p. 2241, September 2022, (in Chinese).[pdf]

3. Yu Tian, D. Song, M. Yang, J. Liu, G. GEeng, M. Zhou, K. Li, and X. Cao, “Uld-net: 3d unsupervised learning by dense similarity learning with equivariant-crop,” J. Opt. Soc. Am. A, 2022, (Accepted). (IF = 2.104) [pdf] [code]

4. J. Liu, D. Song, G. Geng, Yu Tian, M. Yang, Y. Liu, M. Zhou, K. Li, and X. Cao, “Tgps: dynamic point cloud down-sampling of the dense point clouds for terracotta warrior fragments,” Opt. Express, (Accepted). (IF = 3.8) [paper]

A brief cv.