publications

2023

1. 

   FaceLit: Neural 3D Relightable Faces

   Anurag Ranjan, Kwang Moo Yi, Jen-Hao Rick Chang, and Oncel Tuzel

   In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023

   arXiv Code
2. 

   Pointersect: Neural Rendering with Cloud-Ray Intersection

   Jen-Hao Rick Chang, Wei-Yu Chen, Anurag Ranjan, Kwang Moo Yi, and Oncel Tuzel

   In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023

   arXiv Code
3. 

   FastViT: A Fast Hybrid Vision Transformer using Structural Reparameterization

   Pavan Kumar Anasosalu Vasu, James Gabriel, Jeff Zhu, Oncel Tuzel, and Anurag Ranjan

   International Conference on Computer Vision (ICCV), 2023

   arXiv Code

4. 

   FineRecon: Depth-aware Feed-forward Network for Detailed 3D Reconstruction

   Noah Stier, Anurag Ranjan, Alex Colburn, Yajie Yan, Liang Yang, Fangchang Ma, and Baptiste Angles

   International Conference on Computer Vision (ICCV), 2023

   arXiv Code
5. Naturalistic Head Motion Generation from Speech

   Trisha Mittal, Zakaria Aldeneh, Masha Fedzechkina, Anurag Ranjan, and Barry-John Theobald

   2023

   arXiv

2022

1. 

   NeuMan: Neural Human Radiance Field from a Single Video

   Wei Jiang, Kwang Moo Yi, Golnoosh Samei, Oncel Tuzel, and Anurag Ranjan

   In Proceedings of the European conference on computer vision (ECCV), 2022

   Abs arXiv Code

   NeuMan is a NeRF representation of human together with the scene. From a single clip (<100 frames), NeuMan can perform view synthesis of the scene without/with the human in novel poses.

2. 

   MobileOne: An Improved One millisecond Mobile Backbone

   Pavan Kumar Anasosalu Vasu, James Gabriel, Jeff Zhu, Oncel Tuzel, and Anurag Ranjan

   Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022

   Abs arXiv Code

   Fastest neural architecture on iPhone, runs under 1 ms.

3. 

   LCS: Learning Compressible Subspaces for Adaptive Network Compression at Inference Time

   Elvis Nunez, Maxwell Horton, Anish Prabhu, Anurag Ranjan, Ali Farhadi, and Mohammad Rastegari

   Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2022

   Abs arXiv

   LCS is a method for training a "compressible subspace" of neural networks that contains a fine-grained spectrum of models that range from highly efficient to highly accurate. These models require no retraining, thus the subspace of models can be deployed entirely on-device to allow adaptive network compression at inference time.

4. 

   Token Pooling in Vision Transformers

   Dmitrii Marin, Jen-Hao Rick Chang, Anurag Ranjan, Anish Prabhu, Mohammad Rastegari, and Oncel Tuzel

   Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2022

   Abs arXiv

   Token Pooling is a novel downsampling operator for vision transformers that efficiently exploits redundancies in the images and intermediate token representations. Applied to DeiT, it achieves the same ImageNet top-1 accuracy using 42% fewer computations.

5. SPIN: An Empirical Evaluation on Sharing Parameters of Isotropic Networks

   Chien-Yu Lin, Anish Prabhu, Thomas Merth, Sachin Mehta, Anurag Ranjan, Maxwell Horton, and Mohammad Rastegari

   Proceedings of the European Conference on Computer Vision (ECCV), 2022

   arXiv Code

2021

1. 

   Hypersim: A photorealistic synthetic dataset for holistic indoor scene understanding

   Mike Roberts, Jason Ramapuram, Anurag Ranjan, Atulit Kumar, Miguel Angel Bautista, Nathan Paczan, Russ Webb, and Joshua M Susskind

   In Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021

   Abs arXiv Code

   Hypersim is a photorealistic synthetic dataset for holistic indoor scene understanding containing 77,400 images of 461 indoor scenes with detailed per-pixel labels and corresponding ground truth geometry.

2020

1. 

   MorphGAN: One-Shot Face Synthesis GAN for Detecting Recognition Bias

   Nataniel Ruiz, Barry-John Theobald, Anurag Ranjan, Ahmed Hussein Abdelaziz, and Nicholas Apostoloff

   In arXiv, 2020

   Abs arXiv

   MorphGAN can animate any face and control using a 3D rig. It’s one-shot and generalizes to in-the-wild unseen faces.

2. 

   GIF: Generative Interpretable Faces

   Partha Ghosh, Pravir Singh Gupta, Roy Uziel, Anurag Ranjan, Michael J. Black, and Timo Bolkart

   In International Conference on 3D Vision (3DV), 2020

   Abs arXiv Code

   GIF generates realistic face images and animate them with a 3D face rig.

3. 

   Learning Multi-Human Optical Flow

   Anurag Ranjan, David T Hoffmann, Dimitrios Tzionas, Siyu Tang, Javier Romero, and Michael J Black

   International Journal of Computer Vision, 2020

   arXiv

4. 

   Learning to Dress 3D People in Generative Clothing

   Qianli Ma, Jinlong Yang, Anurag Ranjan, Sergi Pujades, Gerard Pons-Moll, Siyu Tang, and Michael J. Black

   In Computer Vision and Pattern Recognition (CVPR), 2020

   Abs arXiv Code

   CAPE is a Graph-CNN based generative model for dressing 3D meshes of human body. It is compatible with the popular body model, SMPL, and can generalize to diverse body shapes and body poses. It is designed to be "plug-and-play" for many applications that already use SMPL. The CAPE Dataset provides SMPL mesh registration of 4D scans of people in clothing, along with registered scans of the ground truth body shapes under clothing.

2019

1. 

   Towards Geometric Understanding of Motion

   Anurag Ranjan

   University of Tuebingen Tuebingen, 2019

2. 

   Attacking Optical Flow

   Anurag Ranjan, Joel Janai, Andreas Geiger, and Michael J Black

   In International Conference on Computer Vision (ICCV), 2019

   Abs arXiv Code

   Deep learning based optical flow methods are vulnerable to adversarial attacks. We show that it is very easy to attack these systems in real world by just placing a small printed patch in the scene.

3. 

   Competitive Collaboration: Joint Unsupervised Learning of Depth, Camera Motion, Optical Flow and Motion Segmentation

   Anurag Ranjan, Varun Jampani, Lukas Balles, Kihwan Kim, Deqing Sun, Jonas Wulff, and Michael J Black

   In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019

   Abs arXiv Code

   Unsupervised learning of several interconnected problems in low-level vision: single view depth prediction, camera motion estimation, optical flow and segmentation of a video into the static scene and moving regions.

4. Unsupervised video segmentation

   Benjamin Ray, and Anurag Ranjan

   Sep 2019

   US Patent 10,402,986

5. 

   Capture, Learning, and Synthesis of 3D Speaking Styles

   Daniel Cudeiro, Timo Bolkart, Cassidy Laidlaw, Anurag Ranjan, and Michael J Black

   In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Sep 2019

   Abs arXiv Code

   A neural network for generating 3D facial motion by using raw speech audio. Works on a veriety of unseen faces.

2018

1. 

   Generating 3D faces using convolutional mesh autoencoders

   Anurag Ranjan, Timo Bolkart, Soubhik Sanyal, and Michael J Black

   In European Conference on Computer Vision (ECCV), Sep 2018

   Abs arXiv Code

   A non-linear model for generating 3D faces using a Convolutional Autoencoder that operates directly on meshes. Our model is state of the art in generating diverse range of 3D facial meshes.

2. 

   Learning human optical flow

   Anurag Ranjan, Javier Romero, and Michael J Black

   British Machine Vision Conference (BMVC), Sep 2018

   Abs arXiv Code

   Learning optical flow for humans is difficult. So, we created a synthetic dataset with realistic humans and trained a neural network on it.

3. Unsupervised learning of multi-frame optical flow with occlusions

   Joel Janai, Fatma Guney, Anurag Ranjan, Michael Black, and Andreas Geiger

   In European Conference on Computer Vision (ECCV), Sep 2018

   Abs PDF Code

   We propose a framework for unsupervised learning of optical flow and occlusions over multiple frames. Our multi-frame, occlusion-sensitive formulation outperforms existing unsupervised two-frame methods and even produces results on par with some fully supervised methods.

2017

1. 

   Optical flow estimation using a spatial pyramid network

   Anurag Ranjan, and Michael J Black

   In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Sep 2017

   arXiv Code
2. Seeing Skin in Reduced Coordinates

   Debanga R Neog, Anurag Ranjan, and Dinesh K Pai

   In IEEE International Conference on Automatic Face & Gesture Recognition (FG), Sep 2017

2016

1. 

   Interactive gaze driven animation of the eye region

   Debanga R Neog, João L Cardoso, Anurag Ranjan, and Dinesh K Pai

   In International Conference on Web3D Technology, Sep 2016

2015

1. Learning periorbital soft tissue motion

   Anurag Ranjan

   University of British Columbia,, Sep 2015

   Abs

   We model the soft tissues around the eyes that are associated with subtle and fast motions and convey emotions during facial expressions. Our data driven model that can efficiently learn and reproduce the complex motion of these periorbital soft tissues.

2. Gaze driven animation of eyes

   Debanga Raj Neog, Anurag Ranjan, João L Cardoso, and Dinesh K Pai

   In ACM SIGGRAPH/Eurographics Symposium on Computer Animation, Sep 2015

   PDF

2012

1. Implementation of 3D object recognition and tracking

   Pankaj Bongale, Anurag Ranjan, and Sahil Anand

   In International conference on Recent Advances in Computing and Software Systems (RACSS), Sep 2012

   PDF