Northwestern Computer Science at CVPR 2026

Northwestern Computer Science researchers showcased a variety of work at the IEEE Computer Society and Computer Vision Foundation 2026 Conference on Computer Vision and Pattern Recognition (CVPR)—from embodied large language models to hyperspectral imaging to training multi-turn agents to an energy-efficient 3D camera inspired by jumping spiders. Held June 3-7 in Denver, CVPR is the preeminent computer vision event for new research in support of AI, computer vision, multi-modal AI, wearable AI, novel AI architectures, spatial computing, agentic AI, AI for science, embodied AI and robotics, and more.

In addition to seven papers presented by participating CS faculty and members of their labs—Emma Alexander (Bio Inspired Vision Lab), Manling Li (Machine Learning and Language Lab), and Han Liu (Modern Artificial General Intelligence and Computer Systems Lab)—Professor Marcelo Worsley was an invited keynote speaker at the Computer Vision for Education (CV4Edu) Workshop.

CVPR 2026 received a record number of 16,092 total paper submissions and about one-quarter—4,089 papers—were accepted to the technical program following peer-review. The following papers featuring Northwestern CS authors were accepted to CVPR 2026:

Bio Inspired Vision Lab

“Spectrum from Defocus: Fast Spectral Imaging with Chromatic Focal Stack”
Team: co-lead authors M. Kerem Aydin and Yi-Chun Hung, both PhD students in computer science; Emma Alexander, assistant professor of computer science; Qi Guo (Purdue University); and Jaclyn Pytlarz (Dolby Laboratories)
Award: Third place for best poster at the 15th IEEE International Workshop on Computational Cameras and Displays
Program: Oral session (top 3 percent of accepted papers)

“SpiderCam: Low-Power Snapshot Depth from Differential Defocus”
Team: co-first authors Marcos Ferreira, a research assistant in the Bio Inspired Vision Lab, and Tianao Li, a PhD student in computer science; Emma Alexander; adjunct associate professor of electrical and computer engineering Josiah Hester, John Mamish, and Yaman Sangar (Georgia Institute of Technology); and Qi Guo (Purdue University)
Program: Paper highlight (top 14 percent of accepted papers)

“Thermal is Always Wild: Characterizing and Addressing Challenges in Thermal-Only Novel View Synthesis”
Team: first author M. Kerem Aydin; Emma Alexander; and Vishwanath Saragadam (University of California, Riverside)
Program: Paper highlight (top 14 percent of accepted papers)

Machine Learning and Language (MLL) Lab

“RAGEN-2 Reasoning Collapse in Agentic RL”
Team: first author Zihan (Zenus) Wang, PhD student in computer science; Manling Li, assistant professor of computer science; Yiping Lu, assistant professor of industrial engineering and management sciences; PhD students in computer science Kangrui Wang, Qineng Wang, and Pingyue Zhang; Shiqi Chen (City University of Hong Kong); Yejin Choi, Li Fei-Fei, and Jiajun Wu (Stanford University); Chi Gui (University of Illinois at Urbana-Champaign); Xing Jin (University of British Columbia); Linjie Li (University of Washington); Licheng Liu (Imperial College London); and Lijuan Wang and Zhengyuan Yang (Microsoft)
Award: Best Paper Award at CVPR workshop on Multimodal Reasoning for Agentic Intelligence

“Learning from Trials and Errors: Reflective Test-Time Planning for Embodied LLMs”
Team: Manling Li and Yejin Choi, Li Fei-Fei, Leonidas Guibas, Yining Hong, Huang Huang, and Jiajun Wu (Stanford University)
Award: Best Paper Award at CVPR workshop on Test-Time Scaling for Computer Vision

“CaptionQA: Is Your Caption as Useful as the Image Itself?”
Team: Manling Li; Emad Barsoum, Zicheng Liu, Ximeng Sun (Advanced Micro Devices, Inc.); Yunong Liu (Stanford University); Chenfeng Xu (University of Texas at Austin); Shijia Yang (Advanced Micro Devices, Inc. and University of Texas at Austin); and Bohan Zhai

Modern Artificial General Intelligence and Computer Systems (MAGICS) Lab

“Towards Sparse Video Understanding and Reasoning”
Team: first author Chenwei Xu, PhD student in statistics; Manling Li; Han Liu, Orrington Lunt Professor of Computer Science and professor of statistics; PhD student in computer science Shang Wu; Zihan Wang; Weijian Li (PhD ’25, MS ’20); former visiting predoctoral scholar Zhuofan Xia; Fan Du, Lie Lu, and Pranav Maneriker (Dolby Laboratories); and Zhen Ye (Johns Hopkins University)

Technological Innovations for Inclusive Learning and Teaching (tiilt) Lab

"Opportunities and Challenges for Multimodal Collaboration Analytics"
Keynote speaker: Marcelo Worsley, Karr Family Associate Professor of Computer Science and associate professor of learning sciences
Program: Computer Vision for Education (CV4Edu) Workshop: Computer Vision × Education: Building a Cross-Community Agenda for Multimodal Vision in Classrooms