Hao (Richard) Zhang (张皓)
GrUVi Lab, School of Computing Science, SFU.
Email: haoz @ sfu . ca
- [new] September 26, 2023: We will be presenting two papers at ICCV 2023 in Paris, two papers at NeurIPS 2023 in New Orleans, and four papers at SIGGRAPH Asia in Sydney, Australia! Congratulations to Maham, Fenggen, Qimin,
Aditya, Yizhi, Akshay, Hang, and Zhiqin, along with all my collaborators.
- [new] August 21, 2023: Please check out our latest survey on Advances in Data-Driven Analysis and Synthesis of 3D Indoor Scenes, in the form of
a Eurographics State-of-the-Art Report (STAR). Great job, Akshay!
- [new] August 1, 2023: I delivered a keynote talk at the Amazon Computer Vision Conference (ACVC) at Palo Alto on August 18.
- [new] June 18, 2023: Join us today for the second workshop on Structural and Compositional Learning on 3D Data at CVPR 2023 in Vancouver.
- [new] June 17, 2023: Congratulations to Akshay and Zhiqin for their successful PhD defenses yesterday. Both theses and defenses received the distinction of "pass as is"! I am proud of you!
- [new] April 6, 2023: I delivered a keynote talk on "An Evolution of Learning Neural Implicit Representations for 3D Shapes" at Computational Visual Media (CVM) 2023.
[new] March 28, 2023: Introducing DS-Fusion (check out the fun video!), a discriminated diffusion based method to automatically generate an artistic typography by stylizing one or more letter fonts to visually convey the semantics of an input word, while ensuring that the output remains readable.
- [new] January 17, 2023: First robotics paper, in ICRA 2023: NIFT: Neural Interaction Field and Template for Object Manipulation. While this work focuses on improving few-shot imitation learning, the really NIFTy idea here, if you are following up on the neural fields literature, is to replace binary occupancies or uni-directional signed distances by a spherical distance function, which is more informative! See Twitter feed.
- [new] January 12, 2023: Welcome Yizhi Wang who is starting his postdoc in 2023. Yizhi obtained his PhD from PKU in 2022 and before arriving in Vancouver, he made a short visit to Shenzhen University and worked there on a collaborative project on neural 3D reconstruction.
- October 18, 2022: Welcome Dingdong Yang who will be starting his PhD in Spring 2023. Dingdong’s latest stint was as a computer vision engineer at Bytedance, the parent company of TikTok. He obtained his MSc degree from the University of Michigan and Bachelor’s degree from Nanjing University.
- September 8, 2022: Nice to see that three of our papers, GRASS, GRAINS, and SDM-Net, are among the list of Top 103 (H5-index) SIGGRAPH/TOG papers in terms of 5-year citations.
- September 3, 2022: Welcome Yilin Liu, a new PhD to join my group from
Prof. Hui Huang's Visual Computing Research Center at Shenzhen University. Yilin comes
to SFU with already three SIGGRAPH (+Asia) papers, one ECCV paper, and a dataset award under his belt!
- August 9, 2002: Check out Neural Dual Contouring (NDC) (SIGGRAPH 2022), a data-driven mesh reconstruction method that is applicable to ALL sorts of inputs including point clouds (without normals as in classical DC), signed or unsigned distance grids, grid of binary voxels, even non-orientable surfaces such as the Mobious strip. It can be a "one-stop" tool for all your surface reconstruction needs, achieving state-of-the-art results! This is collaboration with Google.
- August 7, 2022: A very successful SFU Pre-SIGGRAPH Visual Computing Workshop was held
at SFU downtown campus.
- June 2, 2022: Honored to receive the
Human-Computer Communications Society (CHCCS) Achievement Award in computer graphics. I gave an award keynote at Graphics Interface 2022 in Montreal. See also an Amazon Science Feature about the Award and my role at Amazon.
- May 2, 2022: I am taking a full-time leave for one year to work with Amazon as an Amazon Scholar.
- March 3, 2022: Three papers accepted to CVPR 2022, all on neural implicits:
UNIST (Unpaired Neural Implicit Shape Translation) - aka IM(Implicit)-LOGAN,
RIM-Net (Recursive Implicit Fields for Unsupervised Learning of Hierarchical Shape Structures), and
CAPRI-Net: Learning Compact CAD Shapes with Adaptive Primitive Assembly - aka BSP-Net+++.
while acknowledging Autodesk for their gift fund support!
- February 28, 2022: Congratulations to Manyi who has officially started as an Associate Researcher
at School of Software, Shandong University.
- November 29, 2021: Starting as an Amazon Scholar today.
- November 19, 2021: Talk at the Toronto Geometry Symposium on From Pyramidality to Monotonicity: New Decomposition and Packing Problems for Fabrication and Planning (Youtube video, starting around 13:00). Here is the nice artistically designed talk poster.
- October 16, 2021: Keynote talk on Unsupervised Learning of 3D Shape Structures at the First ICCV Workshop on Structural and Compositional Learning on 3D Data (StruCo3D). Workshop video recording is on Youtube.
- September 24, 2021: Congrats to Zhiqin Chen for winning a
Google PhD Fellowship in Machine Perception, Speech Technology, and Computer Vision! One of two awarded in Canada and the first at SFU CS!
- September 10, 2021: Three papers accepted to SIGGRAPH Asia 2021: Neural Marching Cubes (NMC) learns how to tessellate an implicit filed, given as SDF or binary voxels, from training meshes, TM-Net provides the first part-aware deep generative model for textured meshes, and DronePath presents a tool for continuous aerial path planning for 3D urban scene reconstruction, which is optimized for both reconstruction fidelity and path quality.
- August 31, 2021: Congrats to Akshay for being selected as an Outstanding Reviewer for ICCV 2021!
- August 25, 2021: Former PhD Rui Ma is starting as an associate professor in the
School of Artificial Intelligence at Jilin University, China.
- August 12, 2021: RaidaR will be presented at the Second ICCV Workshop on Autonomous Vehicle Vision (AVVision). To date, it provides the largest annotated image dataset of rainy street scenes to date, to support autonomous driving research: 58,542 rainy images (fog, droplets, road reflections, etc.), 5K with semantic, and 3,658 with instance segmentations. Another dataset paper, 3D-FRONT will appear in ICCV 2021; 3D-FRONT won the inaugural ChinaGraph Best Dataset Award in 2020. Finally, MRGAN has been accepted to the First ICCV Workshop on Structural and Compositional Learning on 3D Data (StruCo3D).
- July 21, 2021: My 3DGV talk on "Compactness, Symmetry, and Functionality: An Evolution to 3D Shape Understanding and Representation" is available on Youtube.
- July 2, 2021: Introducing Neural Marching Cubes (NMC), a data-driven approach for extracting a triangle mesh from a discretized implicit field. NMC recovers sharp edges, a long-standing issue of MC and variants, and reconstructs local mesh topologies more accurately. This work has been conditionally accepted to SIGGRAPH Asia 2021!
- April 19, 2021: Four papers (one oral) to be presented at CVPR 2021:
D^2IM-Net for detail-disentangled single-view 3D reconstruction which can recover surface details,
DECOR-GAN for 3D shape detailization conditioned on a "style code", a structural metric for layout data,
LayoutGMN, that is learned using graph matching networks, and finally
RoofGAN, which learns to generate structured geometry of residential roof structures as a set of roof primitives and their relationships.
Amazingly, all first authors are from SFU. In total, SFU visual computing researchers will present a total of 16 papers in CVPR 2021!
In addition, we will be presenting BalaGAN, an unpaired image-to-image translation network that works effectively over imbalanced domains,
at the L2ID: CVPR Workshop on Learning from Limited and Imperfect Data.
- April 13, 2021: Congrats to Shimon Smith, my MSc student, for being awarded a British Columbia Graduate Scholarship (BCGS) in 2020-21!
- February 22, 2021: I gave an invited talk at UBC CAIDA on learning 3D shape
representations, from implicit functions to structured models.
- December 9, 2020: A preprint about our 3D-FRONT dataset
is available on arXiv. The dataset received the inaugural ChinaGraph Best Dataset Award this year.
- December 4, 2020: Congrats to Zhiqin Chen for being selected as a finalist for
the 2021-22 NVidia Graduate Fellowshop!
- October 29, 2020: Our SIGGRAPH Asia paper on carvable volume decomposition for CNC machining has been selected to appear in the Technical Papers Video Trailer.
- October 24, 2020: I gave a Keynote Talk at ChinaGraph 2020 on "Learning Generative Models of 3D Shapes: From Implicit Functions to Structured Representations".
- October 13, 2020: Very happy to receive the inaugural ChinaGraph Best Dataset Award for our new large 3D indoor scene dataset
Furnished Rooms with layOuts and semaNTics), in collaboration with Alibaba and Chinese Academy of Sciences.
Please consider using it in your next data-driven indoor scene understanding and modeling paper and help us improve it further by providing feedback!
- October 3, 2020: Our COALESCE paper, in collaboration with Adobe, has been accepted to 3DV as an oral. This work develops the first data-driven framework for component-based shape assembly which employs deep learning to synthesize part connections.
- September 29, 2020: Our paper PIE-Net has been accepted to NeurIPS 2020. It is the first deep model to extract parametric curves from point clouds, trained on the ABC dataset. Code will be available very soon.
- September 12, 2020: I will serve on the editorial board of ACM Trans. on Graphcs (TOG) for a three-year term.
- September 7, 2020: An image from our SIGGRAPH 2020 paper "TilinGNN: Learning to Tile with Self-Supervised Graph Neural Network" has been featured on the back cover.
- September 3, 2020: Two papers to appear in
SIGGRAPH Asia 2020: VDAC - volumetric optimization for subtractive manufacturing and
reinforcement learning for transport planning and packing.
- July 4, 2020: Awarded a
Distinguished SFU Professorship (equivalent to University Professor) for a five-year term, 2020-25.
- July 3, 2020: Two papers accepted to ECCV 2020: GANHopper, a multi-hop
GAN for unsupervised image-to-image translation, and
DR-KFS, a differentiable visual similarity metric which improves the structural fidelity
and visual quality of state-of-the-art single-view 3D reconstruction networks. Congrats to Wallace, JC, Johannes, and Akshay!
- June 17, 2020: Our paper BSP-Net: Generating
Compact Meshes via Binary Space Partitioning has won the Best Student Paper Award at CVPR 2020. Congrats to my PhD
Zhiqin Chen, as well as Andrea Tagliasacchi who was
my former PhD and now a senior research scientist at Google. More awards by SFU researchers at CVPR 2020 can be found at
SFU CS News!
- June 16, 2020: Very happy to announce the release
of 3D-FRONT (3D
Furnished Rooms with layOuts and semaNTics), in collaboration with Tao Technology Department of Alibaba Group. 3D-FRONT is a new
large-scale 3D indoor scene dataset containing about
70,000 rooms with layout information, semantic labels, as well as textured 3D furniture objects from 3D-FUTURE. All
are welcome to download and use the dataset. Feedback are welcome!
- June 15, 2020: Congrats to Akshay, whose paper READ: Recursive Autoencoders for Document Layout Generation, has won the Best Paper Award at the CVPR 2020 Workshop on Text and Documents in the Deep Learning Era.
- June 14, 2020: A successful CVPR Workshop on Learning 3D Generative Models. Thanks to all invited speakers, paper contributors, and co-organizers, especially
Daniel Ritchie! Feedback is welcome (Google form from website).
- May 14, 2020: Project page (with talk video, code, etc.)
for our CVPR 2020 oral paper BSP-Net is available. BSP-Net is a differentiable decoder which builds 3D shapes via binary space partitioning and directly outputs low-poly, watertight meshes with sharp details.
- May 14, 2020: Congrats to Kangxue who has just successfully
defined his Ph.D. thesis, "Learning Shape-to-Shape Transformation", and received the honour of "pass as is". Thanks go to the entire thesis and defence
committee: Hui Huang (supervisor), Daniel Cohen-Or
(supervisor), Manolis Savva (internal examiner),
and Angel Chang (chair). Kangxue is my 10-th PhD graduated and he will soon join NVidia Toronto AI Lab
as a research scientist.
- April 30, 2020: Two papers accepted to SIGGRAPH 2020, one on learning to tile
using a self-supervised graph neural network, called TilinGNN (project page), and the other on
Graph2Plan: learning floorplan generation from layout graphs.
- March 11, 2020: Three papers accepted to CVPR 2020, including two orals. Congrats to Zhiqin, Chenyang, and Rundi! First collaborations with Google and with PKU.
- February 15, 2020: Happy to receive a Google Faculty Research Award for 2019-20. Thank you, Google and my Google sponsors!
- February 15, 2020: I am co-organizing two CVPR 2020 workshops: one on Deep Learning Foundations
of Geometric Shape Modeling and Reconstruction and one on Learning 3D Generative Models.
- February 15, 2020: Check out our Eurograph STAR on Learning Generative Models of 3D Structures.
- October 20, 2019: Our paper LOGAN is one of the six papers selected for press release at SIGGRAPH Asia 2019. See a couple of coverages here:
and The News Stack. Please find the code here.
And our paper on LEGO(R) Technic design has been featured in the SIGGRAPH Asia 2019 Technical Paper Trialer.
- September 5, 2019: Check out code and data for our two ICCV 2019 papers:
BAE-NET, a simple three-layer branched AE network architecture that is capable of unsupervised and
one-shot learning for shape co-segmentation, and CompoNet, a deep generative network for shapes via
part synthesis and composition.
- July 29, 2019: Four papers conditionally accepted to SIGGRAPH Asia 2019:
LOGAN for learning general-purpose shape-to-shape transform from unpaired data,
SDM-NET for learning generative models of structured deformable meshes,
RPM-NET for recurrent prediction of motion and parts from raw point scans, and computational LEGO technic design.
- July 23, 2019: Two papers accepted to ICCV 2019: one on branched
autoencoder (BAE-NET), which is extended from IM-NET, for unsupervised and weakly supervised shape co-segmentation,
and another on structured shape synthesis by learning part generation and composition to improve generative diversity.
Congrats to Zhiqin and Nadav!
- July 9, 2019: I gave a keynote talk at SGP 2019 on learing generative models of 3D shapes.
- May 8, 2019: Slides for our Eurographcs tutorial on Learning Generative Models of 3D Structures are available on-line. The presenters include Daniel Ritchie, Kevin Xu, Sid
Chaudhuri, and myself.
- May 1, 2019: Our paper on IM-NET, learning implicit fields for generative modeling of 3D shapes will be presented at CVPR 2019; it has also been invited to be presented at the Workshop on 3D Scene Generation at CVPR.
- March 1, 2019: I am grateful to Adobe for their continued support of my research over the past three plus years via gift funding. It has been a great pleasure to work with all my Adobe collaborators, Sid, Matt, Radomir, Nathan, and Ersin! I look forward to more to come.
- February 19, 2019: Congrats to Chenyang, who successfully defended his
PhD thesis today! His thesis is titled: "Learning to Correspond and Compose Shape Structures". He is now an assistant professor at National University of Defense Technology (NUDT) in China.
- December 6, 2018: One of our latest works GRAINS is able to generate a plausible 3D indoor scene from random noise in fractions of a second. This work has been accepted to ACM TOG and will be presented at SIGGRAPH 2019.
- September 20, 2018: Five SIGGRAPH Asia 2018 papers and one ACM TOG paper (pending minor revision) are available on the papers page.
Very happy for four students Wallace, Akshay, Manyi, Shuhua, and two current and former postdocs Ali and Han, who are "first-timers"! See you in Tokyo for Asia 2018!
- July 20, 2018: Today, officially wrapped up summer travel in 2018. Since late May, I have given 12 talks in 9 cities
including Nanjing, Hangzhou, Lisbon (SMI'18 Keynote), Bintan Island of Indonesia (CGI'18 keynote), Tianjin, Hsingchu, Tainan (Taiwan CGW Keynote), Shenzhen, and Qingdao.
- May 7, 2018: We will present four papers at SIGGRAPH 2018, including one TOG paper. The three SIGGRAPH 2018 papers are
P2P-NET: Bidirectional Point Displacement Net for Shape Transform,
Predictive and Generative Neural Networks for Object Functionality, and
DSCarver: Decompose-and-Spiral-Carve for Subtractive Manufacturing. See publication page for more details.
- March 24, 2018: For the first time, we have an ACM TOG paper
that has been awarded the Graphics Replicability Stamp :) The stamp is "An additional recognition for authors who are
willing to go one step further, and in addition to publishing the paper, provide a complete open-source implementation." Thanks to the first author Fenggen Yu for his
great effort. The paper is on semi-supervised analysis of 3D shape styles and will be presented at
SIGGRAPH 2018 in Vancouver.
- February 21, 2018: SFU CS is launching a NEW Professional Masters Program in Visual Computing in September 2018. It is the first of its kind in Canada. Application deadline is March 30, 2018; see How to Apply.
- November 1, 2017:
We now have a PyTorch implementation of our SIGGRAPH 2017 paper: Generative Recursive Autoencoder for Shape Structures.
- September 19, 2017: We will present a new SIGGRAPH (Asia) course on Modeling and Remodeling 3D Worlds in Thailand on November 29, 2017. Please check out the website periodically for updates.
- September 15, 2017: Rui Ma has successfully defended his PhD thesis on "Sub-Scene Level Analysis and Synthesis of 3D Indoor Scenes". Congrats to Rui!
- September 13, 2017: Two papers accepted to SIGGRAPH Asia 2017:
1) Learning to Predict Part Mobility from a Single Static Snapshot;
2) Learning to Group Discrete Graphical Patterns.
- August 26, 2017: Our paper "ExquiMo: An Exquisite Corpse Tool for Collaborative 3D Shape
Design", first authored by Warunika Ranaweera, has won a Best Paper Award at CAD/Graphics 2017. Congrats, Warunika!
- August 25, 2017: I gave a keynote talk on "Can Machines Learn to Generate 3D Shapes" at the
Biannual International Conference on Computer Aided Design and Graphics (CAD/Graphics) in Zhangjiajie, China.
- July 22, 2017: Our recent paper GRASS (Generative Recursive Autoencoders for Shape Structures)
has been selected as one of six SIGGRAPH 2017 papers featured for press release;
see press release here.
- July 22, 2017: Postdoc Changqing Zou
will take up an Assistant Research Scientist position at the University of Maryland Institute for Advanced Computer Studies (UMIACS).
- July 20, 2017: DualGAN:
Unsupervised Dual Learning for Image-to-Image Translation accepted to ICCV 2017. The key idea is to combine dual learning and generative
adversarial networks for unsupervised generic image-to-image translation. The core idea is identical to that of a recent and concurrent work by Zhu et al. on
Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks.
- June 15, 2017: The IEEE Special Issue on Computational Design and Fabrication, which I co-edited with
Bedrich Benes, Dave Kasic, and Wilmot Li, is available now. The issue consists of four articles from Autodesk, Brown, MIT, and the University of Calgary, covering
a diversity of topics. Check it out here.
- April 19, 2017: Two papers accepted to SIGGRAPH 2017 and one accepted to ACM Transaction on Graphics
and to be presented at SIGGRAPH 2017:
1) Generative Recursive Autoencoder for Shape Structures (GRASS);
2) Deformation-Driven Shape Correspondence via Shape Recognition; and
3) Co-Locating Style-Defining Elements on 3D Shapes.
- December 23, 2016: Congrats to Jaime for successfully defending
his MSC thesis on "Towards Learning of a Joint Geometry-Structure Manifold for Shape Exploration"!
- December 9, 2016: Congrats to Warunika for
successfully defending her MSc thesis on "ExquiMo: An Exquisite Corpse Tool for Co-Creative 3D Shape Modeling". I would
like to assure you that no person, dog, or any other animals were killed as the result of conducting this thesis research.
- October 15, 2016: Gave a talk at
the China-Israel Workshop on Visual Computing and Machine Learning on
"Learning Qualitative Multi-Attribute Organization of Diverse Data Collections", Kunming, China.
- September 12, 2016:
I am on sabbatical from September 2016 to April 2017. I will spend four months at
Stanford University (October 1 - January 31) as a visiting professor at the
- September 7, 2016:
Welcome five new members to the group - Han Liu (postdoc), Akshay Patil (PhD),
Nelson Nauata (MSc), Kiana Mostaghasi (MSc), and Zili Yi (visiting PhD).
- August 11, 2016: Gave an invited talk at
Splunk, Vancouver, on ``Data-Driven Geometric Modeling: Stories from Two Sides".
- August 5, 2016: Gave a keynote talk at Canadian Conference on
Computational Geometry (CCCG) 2016 on ``New Geometry Problems in Computational Design and Fabrication".
- June 14, 2016:
Congrats to Zeinab for successfully defending her MSc thesis and publishing her
first graphics paper! Her paper on Learning 3D Scene Synthesis from Annotated RGB-D Images has been accepted to and will be presented at
SGP 2016, as part of the International Geometry
Summit, in Berlin.
- May 18, 2016: My first service award from SFU: 2016 Dean of
Graduate Studies Awards for Excellence in Leadership. See news article.
- May 10, 2016: I am co-guest editor for an
IEEE Special Issue
on Computational Design and Fabrication Meet Computer Graphics. Submission deadline: September 1, 2016.
- May 1, 2016: My very first half marathon run at the
BMO race. Obviously,
I had a PB (personal best :), clocking in a time of 2:01:54. A super nice day in a super magical atmosphere.
- April 24, 2016: Three papers accepted to SIGGRAPH 2016:
1) Legible Compact Calligrams;
2) Learning How Objects Function via Co-Analysis of Interactions;
3) Connected Fermat Spirals for Layered Fabrication.
More detials can be found on the publication page.
- March 15, 2016: Congratulations to Ibraheem
Alhashim, who is the winner of the
2015 Alain Fournier Ph.D. Dissertation Award. This award is
given to the top computer graphics Ph.D. dissertation defended in a Canadian university. Ibraheem defended his PhD thesis in November
2015 and received the "pass as is" honour given by the thesis committee. His thesis is titled "Topology-Varying Shape Matching and
Modeling". See SFU news story and a nice picture of Ibraheem and his daughter.
- March 13, 2016: First time doing a run race with my wife, a 7-miler (11.2 KM), around Burnaby Lake. See results
- February 1, 2016: Check out our invited paper on
"From inspired modeling to creative modeling". And incidentally, this is my
100-th publication :)
- October 21, 2015: Our lab logo has changed, after a logo contest!
The GrUVi lab is a Union of computer GRaphics and VIsion research.
See details about the logo contest and
here for a gallary of submissions from contestants.
- October 7, 2015: I have been invited this year to serve on the
papers committees of SIGGRAPH 2016, Eurographics 2016, SGP 2016, SMI 2016, GMP 2016, and the editorial boards
of the Visual Computer and the Computer & Graphics journals.
- September 14, 2015: Three papers, decompose-and-pack for 3D printing, deformation-driven topology-varying 3D shape correspondence, and generalized cylinder decomposition, have been accepted
and will be presented at SIGGRAPH Asia 2015, which will be held in Kobe, Japan, Nov. 2-5. More details can be
found on the publications page.
- August 10, 2015: PhD student Rui
Ma won an MITACS Globalink Research Award to conduct research
at Microsoft Research Asia in Beijing, from September 2015 to February 2016.
- August 7, 2015: Prof. Daniel
Cohen-Or (Tel Aviv) and Prof. Hui Huang (SIAT, Shenzhen,
China) have been appointed as adjunct professors
at our school. Both of them have been my long-term collaborators.
Dr. Baining Guo (MSRA) completed his three-year term as adjunct professor at SFU CS.
- August 4, 2015: We have a short paper accepted to Pacific Graphics
2015. The core idea is novel and neat: we propose to measure similarity between two shapes based on sparse
reconstruction of shape descriptors, which is well suited to handle both full and partial (with moderate to significant
amount of missing data) shape similarities. See paper here.
- August 2, 2015: Back from a two-week trip in China. During a span
of 16 days, I visited the cities of Nanjing, Beijing, Qingdao, Shenzhen, Hefei, and Suzhou, giving a total of 7 talks on
- July 5, 2015: Honghua Li and our work on
foldabilizing furniture is
featured on SFU news.
- June 30, 2015: Shuyang Sun
successfully defended his MSc thesis on "Structural Co-Consolidation of Shape Collections". Congrats, Shuyang!
- June 23-24, 2015: Talk on
Shape Compaction for Design and Fabrication at the University of Waterloo and
Autodesk Research. Slides available upon requests.
- June 6, 2015:
Interface 2015 concluded in Halifax. Our excellent line-up of
invited speakers were very well received.
Graphics Interface 2016 will be held in Victoria, BC, from May 31 to June 3.
- June 2, 2015: Some
coverage of our latest work on foldabilizing furniture.
- April 24, 2015:
Congrats to Howard Li for successfully defending his PhD thesis on "Shape Compaction via Stacking and Folding"!
- April 10, 2015:
Two papers from SFU conditionally accepted to SIGGRAPH 2015. More details to follow.
- February 10, 2015:
Check our new book on
A Sampler of Useful Computational Tools for Applied Geometry, Computer Graphics, and Image Processing
- December 31, 2014:
Two papers accepted to Eurographics 2015.
- December 15, 2014:
Our work on approximate pyramidal decomposition is receiving some news and media coverage. I showed how the algorithm can be
utilized to print a 3D Christmas tree more efficiently. See:
Sing Tao Daily (in Chineses).
- December 5, 2014:
Code for Morfit
(an interactive surface reconstruction algorithm published at SIGGRAPH Asia 2014) is
- December 5, 2014:
Our SIGGRAPH Asia 2014 Workshop on Creative Shape Modeling and
Design has concluded. See a photgraph taken at the workshop.
- October 31, 2014:
SIGGRAPH Asia 2014 course
program and schedule.
- October 2, 2014: Two papers conditionally accepted to SIGGRAPH Asia 2014:
Pyramidal decomposition (see project page) and Morfit; see papers.
- August 8, 2014: Former Ph.D. Oliver van Kaick
will start a tenure-track assistant professor position at School of Computer Science, Carleton University in September 2014.
- July 15, 2014: Two awards his year,
an NSERC DAS (Discovery Accelerator Supplement) Award ($40K/year for three years) and
a Faculty of Applied Sciences (FAS) Research Excellence Award (early career).
- June 16, 2014: Letter from the president came today,
officially promoted to full professor :-)
- June 4, 2014: As of June 2014, I will take on the role of
co-Editor-in-Chief for Computer Graphics Forum.
- May 1, 2014: Welcome Jaime Vargas joining the GrUVi lab!
- March 29, 2014:
Two papers accepted to SIGGRAPH 2014: "Topology-Varying 3D Shape Creation via Structural Blending" and
"Organizing Heterogeneous Scene Collection through Contextual Focal Points".
- February 25, 2014:
Our paper on sparse dictionary learning for edit propagation of high-resolution
images and video has been accepted to CVPR, to be held in Columbus, Ohio, in June 2014.
- February 19, 2014:
Our paper on spectral global intrinsic symmetry invariant functions has been
accepted to Graphics Interface, to be held in Montreal in May 2014.
- February 10, 2014:
I gave a talk on Shape Compactions at
Dagstuhl Seminar on New Perspectives on Shape Analysis. Send me a private email if you like to see the slides.
- December 2, 2013: I opened up a weibo on Sina: http://weibo.com/u/3926587854
- November 21, 2013: I gave our SIGGRAPH Asia talk on projective shape analysis; it is a "PSA
that shoots two birds with one stone" ;-) The presentation slides can be
accesssed on the right under "Recent talks".
- November 21, 2013: We have completed our SIGGRAPH Asia 2013 course on structure-aware
shape processing. Slides for my two talks can be found on the right under
"Recent talks". A webpage for the course will be set up by Niloy soon.
- September 18, 2013:
Congrat to Anna, who successfully defended her Master thesis today on
"Active Learning For Semantic Labelling Of Airborne LIDAR Data".
- September 5, 2013:
Welcome new grad students to SFU CS! Here you can get a copy of my
short talk at the grad orientation.
- August 15, 2013:
Our paper on projective shape analysis
has been accepted to SIGGRAPH Asia 2013. The key idea is to perform challenging
3D shape analysis by utilizing knowledge gained from images/projection data.
The dataset used
for the project is available now.
- July 25, 2013:
Source code and data
for our layered analysis of facades via symmetry maximization is available.
- July 22, 2013:
Works on point cloud processing from two of our papers,
point cloud consolidation
and edge-aware resampling (EAR),
will be included in CGAL.
June 6, 2013: The technical program
for SGP 2013 is available. This year's conference features
invited speakers, a
two-day graduate school (July 1-2), and a high-quality papers program. I hope to see many
of you in the beautiful Genova, Italy!
- May 31, 2013: Maybe only for a brief period, our paper on field-guided shape part registration (SIGGRAPH
Asia 2012) is featured on the revamped ACM SIGGRAPH homepage.
- May 21, 2013: Our L1-medial skeleton paper is included in the
SIGGRAPH 2013 technical paper video preview. See the whole preview on
- April 22, 2013:
Four papers accepted to SIGGRAPH 2013 and available from the
papers page: 1) Co-Hierachical
Analysis of Shape Structures; 2) Layered Analysis of Irregular Facades via Symmetry Maximization;
3) Qualitative Organization of Collections of Shapes via Quartet Analysis; and
4) L1-Medial Skeleton of Point Cloud (see
code & data).
- April 18, 2013: I will serve on the SIGGRAPH Asia 2013 (Hong Kong) papers committee.
- April 9, 2013: Talk on "High-Level Geometry Processing: Now and New Challenges", given at Tsinghua University.
Slides available upon requests.
- March 18, 2013:
Source code for our mean-curvature
skeleton algorithm is available. The
method computes a curve skeleton (or an intermediate meso-skeleton) from a mesh
and we believe it is the best-performing mesh skeletonization scheme (quality+speed) to date!
- March 8, 2013:
Check out our new paper on bilateral
maps. Unlike the traditional way of defining local shape descriptors,
which are centered at single feature points, the key new idea is to define
the shape descriptors over regions anchored by a pair of feature points.
- March 1, 2013:
Congrats to Oliver
who received a prestigeous
Azrieli postdoc fellowship as well as a postdoc fellowship from Tel
Aviv University (TAU) in 2013. He
will visit Prof. Danial Cohen-Or at TAU.
- December 20, 2012:
I will serve on the SIGGRAPH 2013 papers committee.
- December 14, 2012:
THE kiss! The best moment of 2012 is that kiss. See picture on
- November 29, 2012:
I am co-chairing SGP 2013 with
Yaron Lipman. SGP 2013 will be held on July 3-5, in Genova, Italy.
- November 26, 2012:
A image from our active
co-analysis paper (to appear at SIGGRAPH Asia 2012 in Singapore) was
chosen as the front image of the
ACM Transaction on Graphics (TOG) proceeding.
- November 1, 2012:
We have received an NSERC Engage Grant in collaboration with
iWonderLearning Inc., working to develop iOS applications that
integrate 3D modeling with language learning. The general theme of the
project is experiential and visual learning.
- October 25, 2012:
We welcome two visiting students supported by the China Scholarship
Council. Ruizhen Hu
arrives from Zhejiang University and Yi Chen is from the
National University of Defense Technology. Both of them start their
visits in Fall 2012 and will stay at the GrUVi lab for a year.
- September 24, 2012:
Final version of our curve style paper is available from the
papers page. This work is a bit unconventional; it
attempts to extract curve styles from a set of shape silhouettes without
a conceptual definition of what the style is. The approach is data-driven
and relies an operational definition for style-content separation.
- September 19, 2012:
Final versions of our four
SIGGRAPH Asia 2012 papers
are now available from the papers page. Also
see the right column of this page. Code and datasets for some of these papers are
also made available, please let me know if you have any comments or suggestions.
- September 10, 2012:
We have made available a dataset, the
Shape COSEG Dataset,
to help researchers evaluate their shape co-segmentation algorithms.
- August 28, 2012:
Welcome three new graduate students: Rui Ma (Ph.D.) from Jilin University,
Shuyang Sun (Ph.D.) from Zhejiang University, and Sushant Joshi (M.Sc.) from
Banaras Hindu University.
- June 29 - July 8, 2012:
I will have an intense 10-day trip to China, visiting MSRA, BUAA, USTC, SIAT, and NUDT, and
giving 6 talks along the way!
- June 29, 2012:
Andrea's paper on
mean curvature skeletons has been accepted to SGP 2012.
The final version of the paper is available from the papers page.
- June 28, 2012:
Nima successfully defended his masters
thesis on 5-6-7 meshes today.
Now he is on his way to work at Facebook.
- May 9, 2012: Our SIGGRAPH 2012 paper on set evolution for 3D shape modeling is available on the
publication page. A short clip of the paper's
video has been selected as part of the SIGGRAPH 2012 Technical Paper Video Trailer!
- I gave a talk at the Center of Scientific Computing (CSC) seminar on
Symmetry Analysis on Shapes and Structures. Since the talk covered
unpublished material, the slides are only available through requests.
- March 14, 2012: Check out our cute
paper on 5-6-7 meshes, which resulted from a course project
by Nima. The paper will be presented at
Graphics Interface 2012 in Toronto in June.
- February 28, 2012: I will serve on the technical
papers committee for SIGGRAPH Asia 2012, to be held in Singpore, November 28 - December 1, 2012.
- February 26, 2012:
Hui's paper on empirical mode decomposition for meshes has been
conditionally accepted to Geometric Modeling and
Processing (GMP) 2012, which will be held in Yellow Mountain in China in June.
See publications page.
- February 22, 2012: Dr. Zhenbao Liu from NWPU of
China is visiting the group for a month. He will conduct collaborative research in the area of shape
analysis and retrieval.
- January 19, 2012: Our five-year NSERC CRD project in collaboration
with Object Raku Technology is starting officially this January. Raku specializes
in LIDAR data processing and analysis technologies. The project is on LIDAR data feature extraction and receives funding
from NSERC and Raku in the amount of $300,000 over five years.
- January 15, 2012: Eric and Kevin, who visited the group
for two years each, have both become Assistant Professors in the National University of Defense Technologies, China.
Kevin also won a Best Thesis Award in his university on "Semantics-Driven 3D Shape Analysis and Modeling".
- December 19, 2011: PPT slides
for two courses I gave at SIGGRAPH Asia 2011 in Hong Kong
are available from the right column: one on "Towards High-Level Geometry Processing" as part of the course on
"Elements of Geometry Processing" with Bruno Levy and one on
"Towards the First SIGGRAPH Paper" as part of the course on "How to Write a SIGGRAPH Paper".
- December 16, 2011: Congrats to
Ibraheem on having his first paper accepted (to the Visual
Computer). See video and paper.
- November 30, 2011:
Oliver van Kaick successfully defended his
Ph.D. thesis on "Matching Dissimilar Shapes". Congrats, Dr. van Kaick!
- November 28, 2011:
I will serve on the SIGGRAPH 2012 papers committee.
- September 20, 2011:
Two of our papers "Unsupervised Co-Segmentation of a Set of Shapes via
Descriptor-Space Spectral Clustering" and "Structure-Preserving Retargeting of Irregular 3D
Architecture" have been accepted to SIGGRAPH Asia 2011. Online versions of
the papers are now available.
- August 23, 2011: I co-organized a five-day Banff
Workshop on "Geometry for Anatomy", August 28 - September 2. See
my talk on symmetry analysis for
shape processing (18 MB).
- August 17, 2011:
Check out some animation pieces submitted
for the final project of CMPT 466 (Computer Animation)
- August 16, 2011:
Oliver van Kaick receives an MITACS
Elevate Postdoc Fellowship valued at
$65,000 per year. He will take up the position with our industrial partner
PDFTron in Vancouver.
- August 12, 2011:
Check out some of the recent paper and invited talks in the right column, including
the SIGGRAPH 2011 talk on photo-inspired modeling and
the talk on Towards
High-Level Geometry Processing I have given several times this year.
- June 27, 2011:
Our VASE (volume-aware surface evolution) paper has been accepted to
SGP 2011. Congrats to
- June 22, 2011:
Slides for SMI 2011 talk on point cloud silhouettes is
available. In retrospect, the paper is really about how to find proper k-nearest neighbors (kNNs)!
- April 7, 2011:
Helmut Pottmann gives
a Distinguished Lecture on Geometric Computing for Freeform Architecture.
- April 6, 2011:
successfully defended his Ph.D. thesis on "A Plane View of Geometric
Silhouettes". Congrats, Matt!
- March 30, 2011: Congrats to
who is awarded an Alexander Bell Canada Graduate Scholarship from NSERC;
the award is given to the most selective group of scholarship
- February 11, 2011:
Congrats to Ibraheem
who successfully defended his MSc thesis on detail-replicating shape
stretching; see demo
video and code.
- December 4, 2010:
Upcoming talk by Steven (Shlomo) Gortler
on December 17. See talk
- November 10, 2010:
Upcoming talk by Ligang Liu who will be visiting me between Nov. 21 and 24. See
- September 13, 2010:
Final version of our three SIGGRAPH Asia papers are now available:
1. cone carving;
2. tree skeleton reconstruction;
3. style-content separation.
- August 20, 2010:
Visit from Johannes Kopf
(Microsoft Research) and Dani
University) and talk by Johannes on Street Slide and Automatic Generation
of Destination Maps.
- July 26, 2010: Our
on spectral mesh processing has concluded. See this
wiki from ALICEpedia for all the course-related information,
including presentation slides. A previous version of the
course was given at SIGGRAPH Asia 2009.
- May 12, 2010: Slides for my Eurographics 2010
Report (STAR) talk on shape correspondence is available here.
- April 30, 2010: Andrea
Tagliassachi has received the MITACS Best
Student Paper Award for 2009. Congrats, Andrea!
- January 12, 2010: One
of our papers won the 2010 Most Cited Paper Award for
the journal Computer-Aided Design.
Papers for this distinction are determined solely based on the highest number
of cites, excluding self-citations, received for all journal articles published in
- November 16, 2009: Due to popular demand, we are making the MATLAB code for
our earlier paper on
robust spectral shape correspondence available. Please download
this ZIP file. The code can be easily adapted to compute
the eigenvalue shape descriptors (EVD) from this paper. Instructions are
in the README file. Note that this code does not contain a Nystrom approximation component,
hence only somewhat small meshes (up to 5K vertices) should be tried on. The code was
written by Varun Jain and last updated by Oliver van Kaick. If you have
any questions, please ask Oliver or myself.