Workshop Papers
Original contributions up to 8 pages (excluding references) following CVPR 2025 author guidelines; reviewed double-blind; not archived in CVPR proceedings.
Computer Vision/CORE Unranked
CViW 2026
Fifth Workshop on Computer Vision in the Wild
Jun 03 2026Official workshop site Site reachable
The 5th Workshop on Computer Vision in the Wild (CVinW) at CVPR 2026 focuses on advancing multimodal AI agents capable of perception, reasoning, and action in dynamic, real-world environments. It emphasizes large multimodal models (LMMs) and vision-language-action (VLA) systems, with key interests in spatiotemporal reasoning, causal inference, long-horizon planning, memory, and robust tool use beyond static image understanding.
Key deadlines
Verified September 9, 2026Full paper
April 21, 2026
AoE
Workshop timeline
Submission and decisions
Full paperKey deadline
April 21, 2026 · AoE
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Challenge Submissions
Submissions for two evaluation challenges: MindCube (Spatial Mental Model Reasoning) and SITE-Bench (Spatial Intelligence Thorough Evaluation).
Research areas in scope
01
Topics of Interest
LMMs & Vision-Language Systems: Open-vocabulary learning, multimodal pretraining, and adaptationVideo Understanding & Temporal Reasoning: Long-range video modeling, causal reasoning, and instruction-followingVLA & Embodied AI: Multimodal action learning, simulation-to-real transfer, and robotic perceptionFoundation Models for Vision Tasks: Object detection, segmentation, tracking, and fine-grained recognition in the wildEfficient Training Methods: Large visual model adaptation methods, measured by #training samples (zero-shot and few-shot), #trainable parameters, throughput, training costNew Metrics and Benchmarks: Novel ways to evaluate existing LMMs and large vision models for task-level transfer and open-set visual recognition
Policies worth checking twice
- Submissions must be original and not previously published.
- All submissions are limited to a maximum of 8 pages (excluding references).
- Submissions must follow the CVPR 2025 author guidelines.
- The review process is double-blind.
- Accepted papers are NOT archived in the CVPR 2026 proceedings.
- Submissions must be made via the Open Review portal.