News
-
[Jan 2025] Our paper on unsupervised, temporally-consistent reconstruction of a time-varying shape (e.g. point clouds from motion-capture), expanding and updating our ICCV 2021 paper, was accepted to TPAMI after spending 3 years (!) in a single review cycle. Fortunately, it is not yet out-of-date :).
-
[Jul 2024] Two papers accepted to ECCV 2024, one on local detailization of 3D shapes, and the other on rig-free motion transfer.
-
[May 2024] Our paper on a neurosymbolic system for visual program inference, with both generative and discriminative applications, was accepted to ICML 2024.
-
[Apr 2024] Our paper on generating 3D shapes via intermediate synthesis of shape skeletons was accepted to SIGGRAPH 2024 (conference track).
-
[Feb 2024] I will serve as an Associate Editor of IEEE Trans. Visualization and Computer Graphics (TVCG).
-
[Sep 2023] I will serve on the SIGGRAPH 2024 Technical Papers Committee and the Eurographics 2024 International Program Committee.
-
[Aug 2023] Our paper on leveraging generative models to explore high-quality mesh deformations was accepted to SIGGRAPH Asia 2023 (conference track).
-
[Mar 2023] Our paper on fast complementary dynamics -- real-time secondary motions orthogonal to rig-induced deformation -- was accepted to SIGGRAPH 2023 (journal track).
-
[Feb 2023] Our paper on unsupervised shape reconstruction by part retrieval and assembly was accepted to CVPR 2023.
-
[Jan 2023] I'm an awardee of the Indian National Science Academy's Medal for Young Scientists, 2022. I'm grateful to the students, collaborators, and mentors who made our research possible.
-
[Jan 2023] I will serve as an Associate Editor of Computer Graphics Forum (2023-2026).
-
[Jul 2022] I will serve on the SIGGRAPH 2023 Technical Papers Committee.
-
[Jul 2022] I will teach a session on geometric deep learning at the ACM India Summer School on Shape Modelling.
-
[Jul 2022] Two papers accepted to ECCV 2022, the first on a patch-based approach for detail-preserving 3D shape completion, and the second on connection-aware reasoning for automatically assembling shapes from parts.
-
[Jun 2022] I participated in a seminar on "Constitutional Rights and Facial Recognition Technology" organized by DSCI and Ashoka University. I discussed the underlying technologies, and the dangers of unregulated deployment (and whether they should be deployed at all). Recording at first link.
-
[Apr 2022] Our paper on neural Jacobian fields for shape mapping was accepted to SIGGRAPH 2022 (journal track).
-
[Mar 2022] Our paper on augmenting geometric shape spaces by alternating between generative model training and energy-guided latent space exploration was accepted to CVPR 2022.
-
[Jan 2022] I will serve on the Program Committees of the IJCAI-ECAI 2022 Special Track on AI, the Arts and Creativity, and SMI 2022.
-
[Aug 2021] Our paper on optimizing the geometry of body-supporting surfaces (e.g. seats of chairs or soles of shoes) to improve comfort, using a novel ergonomic objective based on differentiable physical simulation of a virtual human, was accepted to TVCG.
-
[Jul 2021] Two papers accepted to ICCV 2021, one on a new dataset and learning method for labeling parts of 3D buildings as an oral, and one on temporally-consistent atlas-based surface reconstruction as a poster.
-
[Jun 2021] Our paper on a batch selection policy for active metric learning, that uses submodular optimization of joint entropy to balance informativeness and diversity, was accepted to ECML PKDD 2021.
-
[Jun 2021] I will deliver a keynote talk at Shape Modeling International (SMI) 2021 in November.
-
[Jun 2021] Our CVPR workshop on Learning to Generate 3D Shapes and Scenes will take place on Friday Jun 25.
-
[Apr 2021] Recordings of two recent talks are available online: 3DGV Seminar and Toronto Geometry Colloquium.
-
[Apr 2021] I am co-organizing a workshop on Structural and Compositional Learning on 3D Data at ICCV 2021.
-
[Mar 2021] Two papers accepted to CVPR 2021, one on conditional detailization of shapes as an oral, and one on joint retrieval and deformation as a poster.
-
[Mar 2021] I will deliver talks on 3D geometry processing at the 3DGV Seminar (April 14th India/April 13th US) and the Toronto Geometry Colloquium (April 21st).
-
[Feb 2021] I will serve on the Program Committee of SGP 2021.
-
[Oct 2020] Two papers accepted to 3DV 2020, one on component assembly with joint synthesis as an oral, and one on motion annotation programs as a poster.
-
[Jul 2020] Our paper on approximating 3D point clouds with collections of parametric primitives including b-spline patches was accepted to ECCV 2020.
-
[Jun 2020] The recording of my (virtual) talk on reducing supervision for shape segmentation, at the workshop on Deep Learning Foundations of Geometric Shape Modeling and Reconstruction at CVPR 2020, is now online.
-
[Apr 2020] Our paper on decorrelating batches of triplets for active metric learning was accepted to IJCAI-PRICAI 2020.
-
[Mar 2020] Our paper on neural subdivision was accepted to SIGGRAPH 2020.
-
[Feb 2020] Three papers accepted to CVPR 2020, including two for oral presentation.
-
[Feb 2020] I will serve on the Program Committee of SGP 2020.
-
[Dec 2019] Our state-of-the-art report (STAR) on "Learning Generative Models of 3D Structures" was conditionally accepted to Eurographics 2020.
-
[Dec 2019] I gave a keynote talk on "Learning to Generate 3D Structures" at NCVPRIPG 2019. My slides are now online.
-
[Dec 2019] I am co-organizing a workshop on "Learning 3D Generative Models" at CVPR 2020. Please submit your papers! We also have an exciting lineup of invited speakers.
-
[Dec 2019] I will serve on the Program Committees of IJCAI 2020 and Graphics Interface 2020.
-
[Oct 2019] Thea documentation is now available online via GitHub Pages + Travis.
-
[Jul 2019] Our paper on a new neural network architecture for unsupervised, weakly supervised and one-shot 3D shape segmentation was accepted to ICCV 2019.
-
[Jun 2019] I will give a talk on deep recursive models for scene synthesis at the 3D Scene Generation workshop at CVPR 2019.
-
[May 2019] Our paper on deep metric learning using both distinguishable (orderable) and indistinguishable (unorderable) triplets, to model the perceptual similarity of haptic textures, was accepted to WHC 2019.
-
[Feb 2019] Our paper on a unified deep representation of 3D shapes was accepted to CVPR 2019.
-
[Feb 2019] I will co-teach a tutorial on learning generative models for 3D structures at Eurographics 2019.
-
[Feb 2019] I will give a talk on deep recursive models for shape synthesis at the inaugural NVIDIA AI workshop.
-
[Feb 2019] I will serve on the Program Committees of SGP 2019, IJCAI 2019, and CAD/Graphics 2019.
-
[Oct 2018] I received the IIT Bombay Early Research Achiever Award (one of four recipients) for 2017.
-
[Oct 2018] I will serve on the Program Committee of Graphics Interface 2019.
-
[Sep 2018] Our paper on a recursive neural network architecture for indoor scene generation was conditionally accepted (with minor revisions) to TOG.
-
[Aug 2018] Our paper on shape composition using recursive substructure priors was accepted to SIGGRAPH Asia 2018.
-
[Jul 2018] We posted a new preprint on a recursive neural network architecture for indoor scene generation.
-
[Jul 2018] Our paper on learning local geometric descriptors that are sensitive to the physical material of the object was accepted to 3DV 2018.
-
[Apr 2018] My colleagues at IITB released a whitepaper analyzing the shortcomings of the current Aadhaar national identity scheme, and proposing technical and policy guidelines for a sounder approach. We have now posted a short summary of the whitepaper. All writings on the subject are collected here.
-
[Feb 2018] Our paper on a novel neural network architecture for weakly-supervised shape segmentation was accepted to CVPR 2018.
-
[Jan 2018] Our paper on training deep networks for domain generalization was accepted to ICLR 2018.
-
[Dec 2017] I will give a plenary talk at NCVPRIPG 2017, one of the two major national conferences on visual computing.
-
[Nov 2017] The ACM selected ComplementMe, our label-free part suggestion technique, to feature in a press release.
-
[Oct 2017] I will serve on the Program Committee of Shape Modeling International (SMI) 2018.
-
[Aug 2017] We posted two new preprints: on a novel neural network architecture for weakly-supervised shape segmentation; and on a deep recurrent architecture for cleaning motion capture data.
-
[Jul 2017] Our paper on a weakly-supervised neural network-based method for component suggestion was conditionally accepted to SIGGRAPH Asia 2017.
-
[Jul 2017] The ACM selected GRASS, our generative recursive autoencoder for shape structures, to feature in a press release.
-
[Jun 2017] I will serve on the Program Committee of AAAI 2018 (and also the Demos program).
-
[Mar 2017] One paper on recursive autoencoders for 3D shapes conditionally accepted to SIGGRAPH 2017, and another paper on multi-view CNN-based 3D point descriptors conditionally accepted to TOG.
-
[Feb 2017] Our paper on 3D shape segmentation with view-based CNNs was accepted to CVPR 2017. Camera-ready version with finalized stats coming soon.
-
[Dec 2016] We have a new preprint on 3D shape segmentation with view-based CNNs.
-
[Dec 2016] I taught a tutorial on high-level shape analysis at ICVGIP '16 in IIT Guwahati. My slides are now online.
-
[Jun 2016] Sanjeev Mk won the prestigious Qualcomm Innovation Fellowship 2016 India for his ongoing work on 3D shape analysis. Four such fellowships were awarded, and Sanjeev's was the sole winning team that comprised just one Masters student. Congratulations Sanjeev!
-
[May 2016] Our paper on sketch-driven extraction of customized parts from 3D shapes was accepted to SGP 2016.
-
[Nov 2015] Fuse, the character creator based on my PhD research for which I wrote the first version, hits the big time.
-
[Aug 2015] I moved to the Maximum City.
-
[Apr 2015] Our paper on semantic shape editing with deformation handles (a continuous-deformation counterpart to our UIST 2013 paper) was accepted to SIGGRAPH 2015.
-
[Apr 2015] I was selected as one of five Outstanding Faculty Members across all departments by the Cornell Class of 2018. (Thanks folks!)
-
[Dec 2014] I will co-teach a one-day course on "Data-Driven Visual Computing" at SIGGRAPH Asia 2014 in Shenzhen, with Leo Guibas, Alyosha Efros, Shi-Min Hu, Arik Shamir, Kevin Xu and Jun-Yan Zhu.
-
[Nov 2014] AttribIt was covered in Princeton's Discovery research magazine. "Fierce, Fiercer, Fiercest: Software enables rapid creations -- A new software program makes it easy for novices to create computer-based 3-D models using simple instructions such as 'make it look scarier'."
-
[Aug 2014] Our paper on hierarchical analysis of 3D scenes was conditionally accepted to SIGGRAPH Asia 2014.
-
[Jun 2014] Fuse 1.1 supports adjustment shapes, for freeform deformation of body parts. See the video.
-
[Apr 2014] Our paper on human-centric shape analysis was accepted to SIGGRAPH 2014.
-
[Mar 2014] Fuse is no longer in beta! Mixamo Inc. released version 1.0 of the Fuse character modeling tool at GDC.
-
[Jan 2014] I'm co-organizing the first Tristate Workshop on Imaging and Graphics (TWIG), March 22-23, 2014, at Princeton, co-hosted with (and featuring an extensive list of speakers from) the SIGGRAPH Program Committee meeting.
-
[Jan 2014] Code for the probabilistic model described in our SIGGRAPH '12 paper on component-based shape synthesis is available on the project page. It's offered for reference purposes, as is. If you find a bug in the code, please let us know.
-
[Nov 2013] Mixamo Inc. released an Early Access version of the Fuse character modeling tool on the Steam store. Fuse grew out of my PhD thesis work, and I was the architect and chief developer. Fuse was previously demoed at GDC.
-
[Jul 2013] Our paper on exploratory design with semantic attributes ("make the animal more scary", "make the airplane more military", "make the webpage more artistic") was accepted to UIST 2013.