Adobe Research at ICCV 2021

October 12, 2021

Caption: In this ICCV 2021 paper, researchers developed a model that allows image manipulation based on natural-language text prompts, such as changing the appearance of a cat or a tiger. 

Adobe actively participates in the IEEE Computer Society International Conference on Computer Vision (ICCV) each year.  At this year’s conference, taking place from October 11-17, Adobe is presenting 45 co-authored papers, including 5 oral papers, 34 posters, and 6 workshop papers. Adobe authors have also contributed to the conference in many other ways, including co-organizing several workshops, area chairing, reviewing papers, and giving keynotes at workshops. In addition, Adobe reviewers received several best reviewer awards.

Nearly all of Adobe’s papers are the results of student internships or other collaborations with university students and faculty. For those interested, please check out the Adobe Research Careers website to learn more about internships and full-time career opportunities.

Here are Adobe’s contributions to ICCV 2021.

Technical Papers

A Simple Baseline for Weakly-Supervised Scene Graph Generation 
Jing Shi, Yiwu Zhong, Ning Xu, Yin Li, Chenliang Xu 
 
Adaptive Adversarial Network for Source-free Domain Adaptation 
Taotao Jing, Handong Zhao, Zhengming Ding 
 
AESOP: Abstract Encoding of Stories, Objects, and Pictures 
Hareesh Ravi, Kushal Kafle, Scott Cohen, Jonathan Brandt, Mubbasir Kapadia 
 
ALADIN: All Layer Adaptive Instance Normalization for Fine-grained Style Similarity 
Dan Ruta, Saeid Motiian, Baldo Faieta, Zhe Lin, Hailin Jin, Alex Filipkowski, Andrew Gilbert, John Collomosse 
 
BuildingNet: Learning to Label 3D Buildings 
Pratheba Selvaraju, Mohamed Nabail, Marios Loizou, Maria Maslioukova, Melinos Averkiou, Andreas Andreou, Siddhartha Chaudhuri, Evangelos Kalogerakis 
 
Calibrating Concepts and Operations: Towards Symbolic Reasoning on Real Images 
Zhuowan Li, Elias Stengel-Eskin, Yixiao Zhang, Cihang Xie, Quan Hung Tran, Benjamin Van Durme, Alan Yuille 
 
Contact-aware motion retargeting 
Ruben Villegas, Duygu Ceylan, Aaron Hertzmann, Jimei Yang, Jun Saito 
 
CPFN: Cascaded Primitive Fitting Networks for High-Resolution Point Clouds 
Eric-Tuan Lê, Minhyuk Sung, Duygu Ceylan, Radomir Mech, Tamy Boubekeur, Niloy J. Mitra 
 
Cross-Sentence Temporal and Semantic Relations in Video Activity Localisation 
Jiabo Huang, Yang Liu, Shaogang Gong, Hailin Jin 
 
ECACL: A Holistic Framework for Semi-Supervised Domain Adaptation 
Kai Li, Chang Liu, Handong Zhao, Yulun Zhang, Yun Fu 
 
Editing Conditional Radiance Fields 
Steven Liu, Xiuming Zhang, Zhoutong Zhang, Richard Zhang, Jun-Yan Zhu, Bryan Russell 
 
End-to-End Video Instance Segmentation via Spatial-Temporal Graph Neural Networks 
Tao Wang, Ning Xu, Kean Chen, Weiyao Lin 
 
Face Image Retrieval with Attribute Manipulation 
Alireza Zaeemzadeh, Shabnam Ghadar, Baldo Faieta, Zhe Lin, Nazanin Rahnavard, Mubarak Shah, Ratheesh Kalarot 
 
Feature Importance-aware Transferable Adversarial Attacks 
Zhibo Wang, Hengchang Guo, Zhifei Zhang, Wenxin Liu, Zhan Qin, Kui Ren 
 
Field Convolutions for Surface CNNs 
Thomas Mitchel, Vladimir Kim, Michael Kazhdan 
 
Generative Layout Modeling using Constraint Graphs 
Wamiq Para, Paul Guerrero, Tom Kelly, Leonidas Guibas, Peter 
 
Hierarchical Memory Matching Network for Video Object Segmentation 
Hongje Seong, Seoung Wug Oh, Joon-Young Lee, Seongwon Lee,  Suhyeon Lee, Euntai Kim 
 
HighlightMe: Detecting Highlights from Human-Centric Videos 
Uttaran Bhattacharya, Gang Wu, Stefano Petrangeli, Vishy Swaminathan, Dinesh Manocha 
 
HuMoR: 3D Human Motion Model for Robust Pose Estimation 
Davis Rempe, Tolga Birdal, Aaron Hertzmann, Jimei Yang, Srinath Sridhar, Leonidas J. Guibas 
 
CR-Fill: Generative Image Inpainting with Auxiliary Contextual Reconstruction 
Yu Zeng, Zhe Lin, Huchuan Lu, Vishal M. Patel 
 
Labels4Free: Unsupervised Segmentation using StyleGAN 
Rameen Abdal, Peihao Zhu, Niloy Mitra, Peter Wonka 
 
Language-Guided Global Image Editing via Cross-Modal Cyclic Mechanism 
Wentao Jiang, Ning Xu, Jiayun Wang, Chen Gao, Jing Shi, Zhe Lin, Si Liu 
 
Learning to Cut by Watching Movies 
Alejandro Pardo, Fabian Caba, Juan Léon Alcázar, Ali K. Thabet, Bernard Ghanem 
 
MAAS: Multi-modal Assignation for Active Speaker Detection 
Juan Léon Alcázar, Fabian Caba, Ali K. Thabet, Bernard Ghanem 
 
Modulated Periodic Activations for Generalizable Local Functional Representations 
Ishit Mehta; Michaël Gharbi; Connelly Barnes; Eli Shechtman; Ravi Ramamoorthi; Manmohan Chandraker 
 
Neural Strokes: Stylized Line Drawing of 3D Shapes 
Difan Liu, Matthew Fisher, Aaron Hertzmann, Evangelos Kalogerakis 
 
OSCAR-Net: Object-centric Scene Graph Attention for Image Attribution 
Eric Nguyen, Trung Bui, Vishy Swaminathan, John Collomosse 
 
SemIE: Semantically-aware Image Extrapolation 
Bholeshwar Khurana, Soumya Ranjan Dash, Abhishek Bhatia, Aniruddha Mahapatra, Hrituraj Singh, Kuldeep Kulkarni 
 
SSH: A Self-Supervised Framework for Image Harmonization 
Yifan Jiang, He Zhang, Jianming Zhang, Yilin Wang, Zhe Lin, Kalyan Sunkavalli, Simon Chen, Sohrab Amirghodsi, Sarah Kong, Zhangyang Wang 
 
STEM: An approach to Multi-source Domain Adaptation with Guarantees 
Van-Anh Nguyen, Tuan Nguyen, Trung Le, Quan Hung Tran, Dinh Phung 
 
Stochastic Scene-aware motion prediction 
Mohamed Hassan, Duygu Ceylan, Ruben Villegas, Jun Saito, Jimei Yang, Yi Zhou, Michael J. Black 
 
StyleCLIP: Text-Driven Manipulation of StyleGAN Imagery 
Or Patashnik, Zongze Wu, Eli Shechtman, Daniel Cohen-Or, Dani Lischinski 
 
TEACHTEXT: CrossModal Generalized Distillation for Text-Video Retrieval 
Ioana Croitoru, Simion-Vlad Bogolin, Marius Leordeanu, Hailin Jin, Andrew Zisserman, Samuel Albanie, Yang Liu 
 
Temporally-Coherent Surface Reconstruction via Metric-Consistent Atlases 
Jan Bednarik, Vladimir G. Kim, Siddhartha Chaudhuri, Shaifali Parashar, Mathieu Salzmann, Pascal Fua, Noam Aigerman 
 
Time-Equivariant Contrastive Video Representation Learning 
Simon Jenni, Hailin Jin 
 
Video Pose Distillation for Few-Shot, Fine-Grained Sports Action Recognition 
James Hong, Matthew Fisher, Michaël Gharbi, Kayvon Fatahalian 
 
Weakly Supervised Human-Object Interaction Detection in Video via Contrastive Spatiotemporal Regions 
Shuang Li, Yilun Du, Antonio Torralba, Josef Sivic, Bryan Russell 
 
MVSNeRF: Fast Generalizable Radiance Field Reconstruction From Multi-View Stereo 
Anpei Chen, Zexiang Xu, Fuqiang Zhao, Xiaoshuai Zhang, Fanbo Xiang, Jingyi Yu, Hao Su 
 
Collaging Class-Specific GANs for Semantic Image Synthesis 
Yuheng Li, Yijun Li, Jingwan Lu, Eli Shechtman, Yong Jae Lee, Krishna Kumar Singh

Workshop Papers

Contrastive Feature Loss for Image Prediction 
Alex Andonian, Taesung Park, Bryan Russell, Richard Zhang, Phillip Isola, Jun-Yan Zhu 
Workshop - Advances in Image Manipulation 
 
Defending Object Detection Networks Against Adversarial Patch Attacks 
Thomas Gittings, Steve Schneider, John Collomosse 
Workshop on Adversarial Robustness in the Real World 
 
Learning Where to Cut from Edited Videos 
Yuzhong Huang Xue Bai, Oliver Wang, Fabian Caba, Aseem Agarwala 
Workshop on AI for Creative Video Editing and Understanding 
 
Scene Designer: a Unified Model for Scene Search and Synthesis from Sketch 
Leo Ribeiro, Tu Bui, John Collomosse, Moacir Ponti 
Workshop: Sketching for Human Expressivity 
 
Studying the Effects of Self-Attention for Medical Image Analysis 
Adrit Rao, Jongchan Park, Sanghyun Woo, Joon-Young Lee, Oliver Aalami 
Workshop on Computer Vision for Automated Medical Diagnosis 
 
Seeing the Unseen: Predicting the First-Person Camera Wearer's Location and Pose in Third-Person Scenes 
Yangming Wen, Krishna Kumar Singh, Markham Anderson, Wei-Pang Jan, Yong Jae Lee 
Workshop on Egocentric Perception, Interaction and Computing: Introducing a massive-scale first-person video project

Workshop Co-organizer

StruCo3D Workshop: Structural and Compositional Learning on 3D Data 
Niloy Mitra, Paul Guerrero, Siddhartha Chaudhuri
 
SHE: Sketching for Human Expressivity 
Niloy Mitra 
 
AI for Creative Video Editing and Understanding (CVEU) 
Fabian Caba 
 
LatinX in Computer Vision Workshop (LXCV) 
Fabian Caba 

Invited Talks and Keynotes 

Why Do Line Drawings Work?, at Workshop on Sketching for Human Expressivity 
Aaron Hertzmann 
 
Neural Surface Maps, at Workshop on Deep Learning for Geometric Computing 
Niloy Mitra 
 
Generative Models for Vector Graphics, at Workshop on Unsupervised 3D Learning in the Wild 
Niloy Mitra 

Related Posts