Itzik Ben-Shabat
Itzik Ben-Shabat

The CVPR 2026 Survival Guide: 10 Focused Calendars So You Don't Get Lost in Denver

TL;DR: CVPR 2026 is enormous. 31 workshop tracks, 11 tutorials, and 4,089 papers across five days in Denver. I went through the workshops and tutorials, grouped them into 10 focus topics, and built a recommended program for each. Pick a topic that fits you, subscribe to its calendar, and you have a curated two days of workshops and tutorials on your phone. Each calendar is below with one-click subscribe links and .ics downloads.

Jump to the 10 calendars

Why I made this

When I attended my first few vision conferences as a PhD student, I would walk in with a vague list of "things that sound cool," bump into someone I hadn't seen in a while at the entrance, and by 11am realize I'd already missed the workshop I actually wanted to be at. If that's where you are this year, this post is for you.

Quick context if you are not steeped in this world: CVPR is the IEEE/CVF Conference on Computer Vision and Pattern Recognition, the flagship annual meeting for the computer-vision research community. The 2026 edition runs June 3 to 7 in Denver, Colorado. Past years I have recapped here: CVPR 2024, CVPR 2023, CVPR 2022, CVPR 2019.

CVPR can feel intimidating. It is easy to get lost. The program is the size of a phonebook (younger readers, ask a parent). Workshops and tutorials are great, but there's a real opportunity cost, every workshop you sit in is three other ones you didn't.

What works for me is coming in with a clear plan and schedule, but staying flexible enough to throw it out if I bump into someone interesting at the coffee break. That second part is important. Don't optimize so hard that you skip lunch with a collaborator you haven't seen in two years to catch the second talk of a workshop. The hallway is where most of CVPR actually happens.

This post is the plan I wish I had: a topic-first map of the program, with curated calendars you can subscribe to in one tap.

Disclosure: I work at Roblox on avatars and animation, so I'm biased toward the gaming/character-animation track. I tried to be fair across the others, but if you spot a workshop I missed in your area, ping me on Twitter or Bluesky and I'll add it.

How to use this CVPR 2026 guide

  1. Pick the one or two topics that match what you actually work on. Don't try to do all ten. You can't.
  2. Subscribe to the calendar(s) using the Google Calendar links below, or download the .ics file to import into Apple Calendar or Outlook.
  3. Treat the calendar as a draft, not a contract. Each calendar lists my primary picks (these are the events that end up on the calendar) plus alternates for the same slots, in case something doesn't match your taste.
  4. Build in slack time. Workshops are 3.5–7 hours long. If you sit through every minute, you have nothing left for the posters, the demos, or (most importantly) the people.

Where it all happens: All workshops and tutorials are at the Colorado Convention Center on Wednesday, June 3 and Thursday, June 4. The main conference (orals, posters, keynotes) runs Friday June 5 – Sunday June 7.

The 10 calendars

Each calendar covers workshops and tutorials on Wed June 3 and Thu June 4, 2026, in Mountain Time. Click Add to Google Calendar to subscribe in one tap (a confirmation page in Google Calendar will appear; click 'Add'). For Apple Calendar or Outlook, use Download .ics and double-click the file to import. The Google subscription auto-updates if I edit the events later; the .ics import is a one-shot copy.

Subscribe vs. download: Subscribing via the Google Calendar link keeps the calendar in sync if I update it. Downloading the .ics is a one-shot copy. If in doubt, subscribe.

1. Generative AI & Diffusion Models

From diffusion theory to movie-grade video. The AI-creation power circuit of CVPR.

If you work on image/video generation, this is the densest track of the week. Start the day with the diffusion principles tutorial, deepen with the analytic theory in the afternoon, then ride the wave through personalization and movie-grade production on day two.

Alternates worth knowing about: Accelerated Diffusion Models (Wed AM tutorial), AI for Content Creation (Wed AM workshop), Long Video Creation (Wed PM), Low-Level Vision Frontiers with GenAI (Thu AM), Video Generative Models: Benchmarks (Thu full day).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 3A · Generative Diffusion Modeling · Sat 10:15-11:30 · Bluebird Ballroom
  • Oral 3C · Generative Editing · Sat 10:15-11:30 · Mile High 1A-2A
  • Oral 3D · Multimodal Modeling · Sat 10:15-11:30 · Mile High 3A-4A

Award candidates to look for:

  • #1 A Style is Worth One Code: Unlocking Code-to-Style Image Generation with Discrete ... · Huijie Liu et al.
  • #1 Black-box Membership Inference Attacks on the Pre-training Data of Image-generatio... · Tao Qi et al.
  • #5 Towards Photorealistic and Efficient Bokeh Rendering via Diffusion Framework · Linxiao Shi et al.
  • #3 FUSER: Feed-Forward Multiview 3D Registration Transformer and SE(3)^N Diffusion Re... · Haobo Jiang et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F

Posters to seek out:

  • #44 HoloCine: Holistic Generation of Cinematic Multi-Shot Long Video Narratives · Yihao Meng et al.
  • #93 GeniNav: Generative Model Driven Image-Goal Navigation via Imagination-Guided Cons... · Yuqi Chen et al.
  • #112 IR-HGP: Physically-Aware Gaussian Inverse Rendering for High- Illumination Scenes ... · Qingan Zhang et al.
  • #135 Diffusion-Based Native Adversarial Synthesis for Enhanced Medical Segmentation Gen... · Hongyu Zhang et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

2. Foundation Models, VLMs & Multimodal Reasoning

Where everything is becoming an LLM with eyes. And a body.

The Foundation Models track at CVPR 2026 is huge. I picked four sessions that span the spectrum: grounded retrieval, 3D-language-action models, unified multimodal agents in the wild, and video LLMs.

Alternates worth knowing about: Multimodal Spatial Intelligence (Wed AM), DataMFM (Wed PM), Knowledge-Intensive Multimodal Reasoning (Thu AM), Multi-Modal Reasoning for AI Agents (Thu PM).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 1A · Multimodal Vision · Fri 9:15-10:30 · Bluebird Ballroom
  • Oral 1C · Efficient Reasoning · Fri 9:15-10:30 · Mile High 1A-2A
  • Oral 3D · Multimodal Modeling · Sat 10:15-11:30 · Mile High 3A-4A
  • Oral 4C · Spatial Reasoning · Sat 14:00-15:15 · Mile High 1A-2A

Award candidates to look for:

  • #1 A Style is Worth One Code: Unlocking Code-to-Style Image Generation with Discrete ... · Huijie Liu et al.
  • #2 Adversarial Style Optimization: Enhancing VLM Jailbreaks by GRPO- based Stylistic... · Bingjun Luo et al.
  • #6 ViT^3: Unlocking Test-Time Training in Vision · Dongchen Han et al.
  • #4 NuWa: Deriving Lightweight Class-Specific Vision Transformers for Edge Devices · Ziteng Wei et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F

Posters to seek out:

  • #68 Training One Model to Master Cross-Level Agentic Actions via Reinforcement Learning · Kaichen He et al.
  • #70 EMO-R3: Reflective Reinforcement Learning for Emotional Reasoning in Multimodal La... · Yiyang Fang et al.
  • #146 OddGridBench: Exposing the Lack of Fine-Grained Visual Discrepancy Sensitivity in ... · Tengjin Weng et al.
  • #224 ENC-Bench: A Benchmark for Evaluating Multimodal Large Language Models in Electron... · Ao Cheng et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

3. Video Models & World Models

Models that predict what happens next. Video generation, world models, long-form understanding.

World models have grown from a niche subject in 2024 to one of the dominant themes of CVPR 2026. This program threads four flagship sessions across both days: the new Video World Models workshop, the GigaBrain challenge, geometry-free novel view synthesis, and long-form video understanding.

Alternates worth knowing about: Creative Visual Content (CVEU, Wed AM), World Models in Multi-modal Reasoning tutorial (Wed PM), 4D World Models workshop (Thu PM).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 2D · Spatio-Temporal Reconstruction · Fri 13:00-14:15 · Mile High 3A-4A
  • Oral 5A · Dynamic Perception · Sun 10:15-11:30 · Bluebird Ballroom

Award candidates to look for:

  • #2 Efficiently Reconstructing Dynamic Scenes One D4RT at a Time · Chuhan Zhang et al.
  • #3 FUSER: Feed-Forward Multiview 3D Registration Transformer and SE(3)^N Diffusion Re... · Haobo Jiang et al.
  • #4 Residual Primitive Fitting of 3D Shapes with SuperFrusta · Aditya Ganeshan et al.
  • #5 SmokeSVD: Smoke Reconstruction from A Single View via Progressive Novel View Synth... · Chen Li et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F
  • Session 2 · Fri Jun 5, 16:00 to 18:00 · ExHall A-F

Posters to seek out:

  • #44 HoloCine: Holistic Generation of Cinematic Multi-Shot Long Video Narratives · Yihao Meng et al.
  • #174 Video Generation with Stable Transparency via Shiftable RGB-A Distribution Learner · Haotian Dong et al.
  • #426 VMonarch: Efficient Video Diffusion Transformers with Structured Attention · Cheng Liang et al.
  • #266 U4D: Uncertainty-Aware 4D World Modeling from LiDAR Sequences, Xiang Xu · Alan Liang et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

4. 3D Vision, Reconstruction & Gaussian Splatting

The 3D pipeline: multi-view, 4D dynamic scenes, open-world understanding.

The "3D from Multi-view and Sensors" track has more workshops than any other this year. If your work touches NeRFs, Gaussian Splatting, or any 3D representation, the program below covers spatial reasoning, the SOTA challenge bar, multi-view supervision, and the dynamic-scene 4D frontier.

Alternates worth knowing about: Urban Scene Modeling (full day Wed), OpenSUN3D (Wed PM), 3D Scene Understanding (Thu AM), Image Matching (Thu PM).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 1D · Computational Imaging · Fri 9:15-10:30 · Mile High 3A-4A
  • Oral 2A · 3D Reconstruction · Fri 13:00-14:15 · Bluebird Ballroom
  • Oral 2B · Materials & Lighting · Fri 13:00-14:15 · Four Seasons Ballroom
  • Oral 2C · Gaussian Splatting & Reconstruction · Fri 13:00-14:15 · Mile High 1A-2A

Award candidates to look for:

  • #1 Customized Fusion: A Closed-Loop Dynamic Network for Adaptive Multi-Task-Aware In... · Zengyi Yang et al.
  • #2 Dual Band Thermal Videography: Separating Time-Varying Reflection and Emission Ne... · Sriram Narayanan et al.
  • #3 MetaSpectra+: A Compact Broadband Metasurface Camera for Snapshot Hyperspectral+ ... · Yuxuan Liu et al.
  • #5 Towards Photorealistic and Efficient Bokeh Rendering via Diffusion Framework · Linxiao Shi et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F

Posters to seek out:

  • #39 RnG: A Unified Transformer for Complete 3D Modeling from Partial Observations · Mochu Xiang et al.
  • #74 Stabilizing Feature Geometry in Noisy Pretrained Models for Robust Downstream Tasks · Quanyu Zhang et al.
  • #150 Learning Anchor in Dual Orthogonal Space for Fast Multi-view Clustering · Yalan Qin et al.
  • #272 Generalized-CVO: Fast and Correspondence-Free Local Point Cloud Registration with ... · Ray Zhang et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

5. Embodied AI & Robotics

Agents that act. Physical AI, multi-agent systems, manipulation, lab-to-real.

The 7th Embodied AI Workshop on Thursday is the central event for this community. Wednesday is a build-up day: start with interactive physical AI, then multi-agent robotic systems in the afternoon.

Alternates worth knowing about: World Models Meet Active Sensing (Wed AM), ScaleBot (Wed PM), Agents in Interaction (Thu), Embodied Reasoning Challenge (Thu).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 4B · Embodied & Agentic Intelligence · Sat 14:00-15:15 · Four Seasons Ballroom
  • Oral 5C · Geometry and Robotics · Sun 10:15-11:30 · Mile High 1A-2A

Award candidates to look for:

  • #2 NitroGen: An Open Foundation Model for Generalist Gaming Agents, Loïc Magne · Anas Awadalla et al.
  • #3 PAI-Bench: A Comprehensive Benchmark For Physical AI · Fengzhe Zhou et al.
  • #4 RefAV: Towards Planning-Centric Scenario Mining · Cainan Davidson et al.
  • #5 SoccerMaster: A Vision Foundation Model for Soccer Understanding · Haolin Yang et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F

Posters to seek out:

  • #68 Training One Model to Master Cross-Level Agentic Actions via Reinforcement Learning · Kaichen He et al.
  • #70 EMO-R3: Reflective Reinforcement Learning for Emotional Reasoning in Multimodal La... · Yiyang Fang et al.
  • #89 Do You Have Freestyle? Expressive Humanoid Locomotion via Audio Control · Zhe Li et al.
  • #93 GeniNav: Generative Model Driven Image-Goal Navigation via Imagination-Guided Cons... · Yuqi Chen et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

6. Autonomous Driving & Self-Driving Systems

WAD, V2X, world-model sims, and the deployment of foundation models on cars.

The Workshop on Autonomous Driving (WAD) is the flagship and traditionally where the big self-driving labs show their hand. The GenAI-simulation tutorial in the afternoon is a great practical complement and ends the day with an actionable sim pipeline.

Alternates worth knowing about: Foundation Models for V2X (full day Wed), Open-World Perception + LLMs for Driving (full day Wed), Visual Odometry workshop (Thu AM).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 1B · Visual Security · Fri 9:15-10:30 · Four Seasons Ballroom

Award candidates to look for:

  • #4 LDP-Slicing: Local Differential Privacy for Images via Randomized Bit-Plane Slicing · Yuanming Cao et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F
  • Session 2 · Fri Jun 5, 16:00 to 18:00 · ExHall A-F

Posters to seek out:

  • #281 DENALI: A Dataset Enabling Non-Line-of-Sight Spatial Reasoning with Low-Cost LiDARs · Nikhil Behari et al.
  • #338 ActiveAD: Planning-Oriented Active Learning for End-to-End Autonomous Driving · Han Lu et al.
  • #344 GuideFlow: Constraint-Guided Flow Matching for Planning in End- to-End Autonomous ... · Lin Liu et al.
  • #118 HG-Lane: High-Fidelity Generation of Lane Scenes under Adverse Weather and Lightin... · Daichao Zhao et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

7. Gaming: Avatars, Motion & Character Animation

If you ship characters, animations, or game worlds, this is your two days.

This is the calendar I built for myself. If you work on avatars, character animation, rigging-adjacent topics, motion capture, or virtual humans, CVPR 2026 has an unusually deep program for you. Day one is the motion-and-mocap stack; day two pulls in human-centric generative models and appearance.

Alternates worth knowing about: Photorealistic 3D Head Avatars (Wed AM), Physical Appearance Modeling & Relighting (Wed AM), CV for Biomechanics (Wed PM), Multimodal Human Motion Analysis (Wed PM), PhysHuman (Thu AM), Appearance Understanding and Generation (Thu PM).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 2B · Materials & Lighting · Fri 13:00-14:15 · Four Seasons Ballroom
  • Oral 5D · Human-Centric Modeling & Lighting · Sun 10:15-11:30 · Mile High 3A-4A

Award candidates to look for:

  • #2 Natural Human Motion Recovery by Aligning High-Order Temporal Dynamics from Monocu... · Dingkun Wei et al.
  • #4 SAM 3D Body: Robust Full-Body Human Mesh Recovery · Xitong Yang et al.
  • #1 3DReflecNet: A Large-Scale Dataset for 3D Reconstruction of Reflective, Transparen... · Zhicheng Liang et al.
  • #2 GLINT: Modeling Scene-Scale Transparency via Gaussian Radiance Transport · Youngju Na et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F
  • Session 2 · Fri Jun 5, 16:00 to 18:00 · ExHall A-F

Posters to seek out:

  • #206 Geometric Neural Distance Fields for Learning Human Motion Priors · Zhengdi Yu et al.
  • #369 Vanast: Virtual Try-On with Human Image Animation via Synthetic Triplet Supervision · Hyunsoo Cha et al.
  • #657 OMG-Bench: A New Challenging Benchmark for Skeleton- based Online Micro Hand Gestu... · Haochen Chang et al.
  • #629 ResiHMR: Residual-Limb Aware Single-Image 3D Human Mesh Recovery for Individuals w... · Jiaying Ying et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

8. Trustworthy AI: Safety, Explainability, Adversarial & Ethics

Things you can defend. To a reviewer, to a regulator, to a user.

This one cleanly tells a story across two days: from threats (Wed AM tutorial), to mitigations (Wed PM unlearning), to interpretability and "how do these models actually work" (Thu full day).

Alternates worth knowing about: Tom Builds, Tom Breaks VLM attacks tutorial (Wed AM), Principled Interpretability tutorial (Wed PM), Synthetic & Adversarial Forensics (Wed PM), Trustworthy/Robust Visual Intelligence (Thu AM), Safe AI for All Domains (Thu AM).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 1B · Visual Security · Fri 9:15-10:30 · Four Seasons Ballroom
  • Oral 5B · Generalization and Adaptation · Sun 10:15-11:30 · Four Seasons Ballroom

Award candidates to look for:

  • #2 Adversarial Style Optimization: Enhancing VLM Jailbreaks by GRPO- based Stylistic... · Bingjun Luo et al.
  • #1 Black-box Membership Inference Attacks on the Pre-training Data of Image-generatio... · Tao Qi et al.
  • #2 Data Leakage Detection and De-duplication in Large Scale Geospatial Image Datasets · Yeshwanth Kumar Adimoolam et al.
  • #4 LDP-Slicing: Local Differential Privacy for Images via Randomized Bit-Plane Slicing · Yuanming Cao et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F

Posters to seek out:

  • #74 Stabilizing Feature Geometry in Noisy Pretrained Models for Robust Downstream Tasks · Quanyu Zhang et al.
  • #135 Diffusion-Based Native Adversarial Synthesis for Enhanced Medical Segmentation Gen... · Hongyu Zhang et al.
  • #263 PRISM: Prototype-based Reasoning with Inter-modal Semantic Mining for Interpretabl... · Anni Yu et al.
  • #300 A Debiased Reconstruction-based Framework for Training- Free Detection of AI-Gener... · Sungik Choi et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

9. Medical & Scientific Vision

Where vision crosses into the wet lab and the clinic.

The medical track is unusually unified this year. Foundation models for medical imaging show up in three different workshops. The 12th Medical Computer Vision workshop on Thursday is the flagship; the Wednesday PHAROS factory session is the most policy-relevant.

Alternates worth knowing about: Multimodal Foundation Models for Biomedicine (Wed AM), CV4Animals (Wed AM), Microscopy Image Analysis (Thu full day), Medical Reasoning with VL FMs (Thu AM), CV4Science (Thu PM), Foundation Models for Medical Vision (Thu PM).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 4D · Visual Segmentation · Sat 14:00-15:15 · Mile High 3A-4A
  • Oral 6C · Medical Vision · Sun 14:00-15:15 · Mile High 1A-2A

Award candidates to look for:

  • #4 R^2-Seg: Training-Free OOD Medical Tumor Segmentation via Anatomical Reasoning and... · Shuaike Shen et al.
  • #2 Learning Diffeomorphism for Medical Image Registration with Time-Embedded Architec... · Mohammadjavad Matinkia et al.
  • #3 Dual-level Adapter Boosting Prompt-free Curvilinear Structure Segmentation · Kai Zhu et al.
  • #5 Medic-AD: Towards Medical Vision-Language Model's Clinical Intelligence · Woohyeon Park et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F
  • Session 2 · Fri Jun 5, 16:00 to 18:00 · ExHall A-F

Posters to seek out:

  • #131 AD-GBC: Anisotropic Granular-Ball Skip-Connection Refiner for UNet- Based Medical ... · Xiya Shen et al.
  • #135 Diffusion-Based Native Adversarial Synthesis for Enhanced Medical Segmentation Gen... · Hongyu Zhang et al.
  • #651 CMR-RD: Long-Tailed Adaptive VLM for Explainable CMR Diagnosis · Yansong Li et al.
  • #125 CROWn: A Unified Framework for Anti-Aliased Downsampling and Phase-Calibrated Fusi... · Xingru Huang et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

10. Efficient, Edge & On-Device Vision

Vision where the compute is cheap and the battery is small.

This track is almost entirely on Wednesday, five workshops in one day. The morning tutorial gives you the practical foundations; FedVision is the privacy-on-device complement; EDGE in the afternoon takes you into generative models on phones.

Alternates worth knowing about: On-Sensor Vision (Wed AM), 22nd Embedded Vision Workshop (Wed AM), Neural Architecture Search (Wed PM), Efficient Deep Learning for CV (full day Wed), Mobile AI workshop (full day Wed).

On the main conference floor (Fri Jun 5 to Sun Jun 7)

Talks

Oral sessions to circle:

  • Oral 1C · Efficient Reasoning · Fri 9:15-10:30 · Mile High 1A-2A
  • Oral 3D · Multimodal Modeling · Sat 10:15-11:30 · Mile High 3A-4A
  • Oral 6D · Large-Scale Neural Modeling · Sun 14:00-15:15 · Mile High 3A-4A

Award candidates to look for:

  • #4 NuWa: Deriving Lightweight Class-Specific Vision Transformers for Edge Devices · Ziteng Wei et al.
  • #5 Plant Taxonomy Meets Plant Counting: A Fine-Grained, Taxonomic Dataset for Counti... · Jinyu Xu et al.
  • #6 Rethinking Dataset Distillation: Hard Truths about Soft Labels, Priyam Dey · Aditya Sahdev et al.
  • #18 Rethinking Dataset Distillation: Hard Truths about Soft Labels · Priyam Dey et al.
Posters

Poster sessions to attend:

  • Session 1 · Fri Jun 5, 10:45 to 12:45 · ExHall A-F

Posters to seek out:

  • #166 Scalable Trajectory Generation for Whole-Body Mobile Manipulation · Yida Niu et al.
  • #312 FedSDR: Federated Graph Learning with Structural Noise Detection and Reconstruction · Jiaqi Liu et al.
  • #403 SSM-Aware Token-Efficient VMamba via Adaptive Patch Pruning and Merging for Person... · Huiyuan Huang et al.
  • #536 LiteSense: Lifting Lightweight ToF with RGB for High-Resolution Metric Depth Estim... · Yusheng Li et al.

The topic calendar above has every flagged paper (orals, highlights, and award candidates) with the exact day and room, auto-synced if the program updates.

Add to Google Calendar · Download .ics

Opening ceremony

Welcome & Awards · Fri Jun 5, 8:30 to 9:00 · Bluebird Ballroom. Kicks off the main conference. Worth being on time for — it sets the tone for the week, and the awards-candidate-tagged papers in each topic calendar are the shortlist for the prizes announced later in the conference.

Main conference keynotes

Three plenary keynotes, all in the Bluebird Ballroom. These tend to be the most-attended single sessions of the conference; show up early.

  • Fri Jun 5, 14:45 to 15:45 · Simon Kohl, Founder & CEO, Latent Labs · Programmable Biology: Generative AI for Molecular Design
  • Sat Jun 6, 10:30 to 11:30 · Jerry Chow, IBM TJ Watson, CTO Quantum Centric Supercomputing · Transforming Computing with Quantum-Centric Supercomputing
  • Sun Jun 7, 10:30 to 11:30 · Thomas Serre, Professor of Cognitive & Psychological Sciences, Brown · Scaling Laws vs. Neural Laws: Toward More Natural Artificial Vision

Best-paper winners are announced during the conference; the awards-candidate-tagged papers in each topic calendar are the shortlist.

CVPR 2026 survival rules

After about ten years going to this conference, the rules I keep coming back to:

  • Plan, but stay flexible. I write down 2 or 3 things I want to see per slot and let myself drop the plan if a hallway conversation gets interesting. The plan is a safety net, not a contract.
  • Pick workshops by the speakers, not the title. A workshop is only as good as its invited talks. Check the speaker list before you commit four hours of your day. A great workshop with strong speakers beats almost anything else on the program; a mediocre one is the most expensive way to be in the wrong room.
  • If you are a PhD student, apply to the Doctoral Consortium next year. It is a small, friendly venue where the people you meet are likely to be your colleagues for the next decade. Note that it is not drop-in: you need to register and be accepted in advance. If you missed the deadline this year, the other strong on-ramp is the affinity workshops (LatinX in CV, Women in CV) and the Wednesday community sessions. (For more on the PhD years in general, I wrote up my advisor-hunt guide.)
  • Take a real break. Conferences are exhausting. Find somewhere scenic, sit down, and have lunch with people you just met. It is the part of the day that pays you back the most, both for your energy and for the new connections. Denver is a mile high, so hydrate before you regret it.
  • Sponsors usually hand out swag. T-shirts, socks, stickers, the occasional rubber duck. I have never actually taken any home, but if you collect that sort of thing, the expo floor is the place.
  • Find the affinity workshops. If you are new, LatinX in CV and Women in CV on Wednesday are some of the friendliest rooms at the conference.

The thing nobody puts in the survival guide

You can read the papers from home. You really can. The PDFs are free, the code is on GitHub, the talks end up on YouTube within a few weeks. None of that is why you flew to Denver.

You flew here for the people. The five-minute conversation at a poster that turns into a Slack thread that turns into a collaboration. The dinner with a group you only ever see at conferences. The PhD student whose work overlaps with yours in a way you would never have noticed online. Long-lasting research collaborations with people you actually like are very hard to build in isolation. A conference is one of the rare places where you can build them on purpose.

So talk to people. Ask the speaker the question afterwards. Walk over to the poster and stay for ten minutes, not two. If you bump into someone interesting at the coffee break and you have a workshop in five minutes, go to the coffee. Most years, that is the part of the conference I actually remember.

Share this

If this was useful, here are some pre-baked share blurbs. Stealing them is encouraged.

LinkedIn: I made 10 focused calendars for CVPR 2026, one per topic (Generative AI, VLMs, World Models, 3D, Embodied, Driving, Gaming, Trustworthy AI, Medical, Efficient Vision). One-click subscribe so you don't get lost in Denver. → https://itzikbs.com/go/cvpr-2026-survival-guide

Twitter / X: CVPR 2026 is huge. I picked through every workshop & tutorial and built 10 topic calendars you can subscribe to in one tap. Pick yours, show up. → https://itzikbs.com/go/cvpr-2026-survival-guide

Bluesky: 10 curated CVPR 2026 calendars, one per topic, generative AI, VLMs, world models, 3D, embodied, driving, gaming, trustworthy AI, medical, efficient. Subscribe in one tap. https://itzikbs.com/go/cvpr-2026-survival-guide

Frequently asked questions

When and where is CVPR 2026?

CVPR 2026 runs June 3 to June 7, 2026 at the Colorado Convention Center in Denver, Colorado. Wednesday June 3 and Thursday June 4 are the workshop and tutorial days. The main conference (oral presentations, posters, keynotes, and the expo) runs Friday June 5 through Sunday June 7.

How do I subscribe to one of the topic calendars?

Click the Add to Calendar button next to any of the 10 topics above. Google Calendar will ask you to confirm a new calendar subscription that auto-updates if I edit the events later. On Apple Calendar or Outlook, use the Download .ics link and double-click the file to import.

I am attending CVPR for the first time. Where should I start?

Start at the top of this post. Skim the 10 topic calendars, find the one that matches what you actually work on, hit Add to Google Calendar, and you have a real schedule for the workshop and tutorial days. From there, plan two or three things per slot and leave room for hallway conversations. Most people remember the people they met better than the talks they sat through.

What is the difference between orals, highlights, and award candidates?

Of the 4,089 accepted papers, the program chairs selected 141 orals (3.4%, organized into four parallel tracks) for live talks, 578 highlights (14.1%) that are flagged in the poster sessions as especially worth visiting, and nominated 74 best-paper award candidates. The calendars above tag each session with the topic-relevant award candidates and orals so you do not have to scan the full program by hand.

Are the CVPR 2026 workshop and tutorial recordings available?

It is not standardized. Some recordings end up on the CVPR Computer Vision Foundation YouTube channel; others stay on the workshop or tutorial homepage; some are never posted publicly at all. Your best move is to check the workshop or tutorial homepage (each calendar above links to it) a few weeks after the conference. Slides are usually easier to find than video.

My research area is not exactly one of the 10 calendars. What now?

The 10 topics are coarse buckets, not a taxonomy. The strongest move is to open the CVPR 2026 program PDF and do a keyword search for your terms — workshops, tutorials, and papers all turn up. Even better, identify the three or four authors who publish closest to your area, search the program for their names, and check what they are presenting and when. Conferences are most useful when you arrive with a short list of specific people you want to talk to.

Spot a CVPR 2026 workshop I missed?

Ping me on Twitter, Bluesky, or LinkedIn and I'll update the post and the calendars. See you in Denver.