Guocheng Qian

钱国成

Ph.D. Candidate in Computer Science

King Abdullah University of Science and Technology (KAUST)

Gordon Guocheng Qian

Hi welcome! I am a research scientist at Snap Research working with Kfir Aberman on Personalized AI Generation. I graduated as a Ph.D. in Computer Science at KAUST, under the supervision of Prof. Bernard Ghanem. During my Ph.D. studies, I was fortunate to intern at serveral leading research institutions, including Snap Research, Meta Reality Lab, Microsoft Research, Megvii Research, and SenseTime Research. Prior to that, I received my B.Eng degree with the highest undergraduate honor from Xi'an Jiaotong University (XJTU). My primary research interests lie in 3D perception and 3D generation. My representative work includes 3D Foundation Model PointNeXt (NeurIPS), ASSANet (NeurIPS Spotlight), DeepGCNs (T-PAMI), and 3D generation Magic123 (ICLR) and AToM.
If you are interested in working in personalized/3D-related generation with us, please drop me a message through guocheng.qian [at] outlook.com

Education

Ph.D. in CS
KAUST , 2019 - 2023
B.Eng in ME (3.9/4.3)
XJTU , 2014 - 2018
Exchange Student
HKUST , 2017 Spring

Interests

3D Perception

PointNeXt, DeepGCNs

3D Generation

Magic123, AToM

Efficient/Mobile Deep Learning

ASSANet

News

2024/02: First-authored paper AToM is preprinted at arXiv.
2024/01: First-authored paper Magic123 get accepts to ICLR.
2023/12: I join Snap Inc. as a research scientist working on personalized generation with Kfir Aberman.
2023/10: ZeroSeg gets accepted to ICCV.
2023/05: I join Snap Inc. as a research scientist intern working on text/image-to-3D.
2022/11: One paper gets accepted to Adavanced Science (IF 2021: 17.52).
2022/10: First-authored paper PointNeXt gets accepted by NeurIPS'22.
2022/08: I joined On Device AI team under XRTech in Meta as an AI research scientist intern.
2022/08: First-authored paper Pix4Point is preprinted at arXiv.
2022/06: First-authored paper TENet gets accepted to ICCP'22.
2022/04: First-authored Paper TNAS gets accepted to CVPR'22 workshop.
2021/10 : First-authored paper ASSANet gets accepted to NeurIPS'21 as a spotlight paper.
2021/03 : Co-first authored paper DeepGCNs gets accpected to journal TPAMI'21 .
2021/03 : First authored paper PU-GCN gets accepted to CVPR'21
2020/12 : I graduate as a Master in Computer Science!
2020/03 : 1 Co-first authored paper SGAS (Sequential Greedy Architecture Search) gets accepted to CVPR’20.
2020/01 : I serve as Teaching Assistant for course: CS390D Deep Learning (2020 Spring)
2018/07 : I graduate from Xi’an Jiaotong University with the highest undergraduate honor (with GPA 3.9/4.3).
2018/06 : I join Sensetime Research as a research scientist intern, supervised by Jimmy S. Ren .

Selected Publications

Check my full publication at Google Scholar

AToM: Amortized Text-to-Mesh using 2D Diffusion

Guocheng Qian, Junli Cao, Aliaksandr Siarohin, Yash Kant, Chaoyang Wang, Michael Vasilkovsky, Hsin-Ying Lee, Yuwei Fang, Ivan Skorokhodov, Peiye Zhuang, Igor Gilitschenski, Jian Ren, Bernard Ghanem, Kfir Aberman, Sergey Tulyakov

February 2024 arXiv'24

AToM trains a single text-to-mesh model on many prompts using 2D diffusion without 3D supervision, yileds high-quality textured meshes under a second, and generalizes to unseen prompts.

Preprint Code Project

Magic123: One Image to High-Quality 3D Object Generation Using Both 2D and 3D Diffusion Priors

Guocheng Qian, Jinjie Mai, Abdullah Hamdi, Jian Ren, Aliaksandr Siarohin, Bing Li, Hsin-Ying Lee, Ivan Skorokhodov, Peter Wonka, Sergey Tulyakov, Bernard Ghanem

January 2024 ICLR'24

Magic123 is a coarse-to-fine image-to-3D pipeline that produces high-quality high-resolution 3D content from a single unposed image by the guidance of both 2D and 3D priors.

Preprint Code Project

Pix4Point: Image Pretrained Standard Transformers for 3D Point Cloud Understanding

Guocheng Qian, Abdullah Hamdi, Xingdi Zhang, Bernard Ghanem

December 2023 3DV'24

We propose a state-of-the-art Standard Transformer model for point cloud understanding and find that image pretraining helps point cloud tasks.

Preprint Code

Exploring Open-Vocabulary Semantic Segmentation from CLIP Vision Encoder Distillation Only

Jun Chen, Deyao Zhu, Guocheng Qian, Bernard Ghanem, Zhicheng Yan, Chenchen Zhu, Fanyi Xiao, Mohamed Elhoseiny, Sean Chang Culatana

June 2023 ICCV'23

ZeroSeg trains open-vocabulary zero-shot semantic segmentation models using only CLIP Vision Encoder

Preprint Code

PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategies

Guocheng Qian, Yuchen Li, Houwen Peng, Jinjie Mai, Hasan Abed Al Kader Hammoud, Mohamed Elhoseiny, Bernard Ghanem

August 2022 NeurIPS'22

PointNeXt boosts the performance of PointNet++ to the state-of-the-art level with improved training and scaling strategies.

Preprint Code

Rethinking Learning-based Demosaicing, Denoising, and Super-Resolution Pipeline

Guocheng Qian, Yuanhao Wang, Jinjin Gu, Chao Dong, Wolfgang Heidrich, Bernard Ghanem, Jimmy S. Ren

August 2022 ICCP'2022

Preprint Code Dataset

ASSANet: An Anisotropic Separable Set Abstraction for Efficient Point Cloud Representation Learning

Guocheng Qian, Hasan Abed Al Kader Hammoud, Guohao Li, Ali Thabet, Bernard Ghanem

October 2021 NeurIPS'21, Spotlight

ASSANet makes PointNet++ faster and more accurate.

Preprint Code

DeepGCNs: Making GCNs Go as Deep as CNNs

Guocheng Qian*, Guohao Li*, Matthias Müller*, Itzel C. Delgadillo, Abdulellah Abualshour, Ali Thabet, Bernard Ghanem

March 2021 TPAMI'2021

This work transfers concepts such as residual/dense connections and dilated convolutions from CNNs to GCNs in order to successfully train very deep GCNs.

Preprint Code Project Slides Video