CVPR 2026 Workshops | SAUAFG: Skilled Activity Understanding, Assessment and Feedback Generation

Introduction

Imagine a world where computer vision-based systems can analyze a video of an athlete, a surgeon, a patient, or a factory worker and instantly provide expert-level actionable feedback---correcting techniques, identifying inefficiencies, and helping people refine their skills in real time. Thanks to rapid progress in video understanding, this vision is becoming reality. AI-powered systems can now analyze complex human activities, assess performance, and generate intelligent feedback, unlocking new possibilities in sports, healthcare, manufacturing, education, rehabilitation, and beyond. Through Expert Keynotes and Invited Contributions, this CVPR 2026 workshop will explore the cutting edge of skilled activity understanding, assessment, and feedback generation, bridging research and real-world applications.

As AI systems become more capable of understanding human expertise, the implications are profound---empowering individuals with personalized coaching, democratized skill development, and scalable training solutions. We invite researchers, industry leaders, and practitioners to join us in shaping the future of AI-powered skill understanding. Whether working on foundational research, applied solutions, or real-world deployment, this workshop is an opportunity and forum to learn about and push the boundaries of how AI perceives, evaluates, and enhances human ability.

Date & Timings: June 04, 2026 | 13:00-18:00 (Afternoon Session)

Venue: Hall 705/707, Denver Convention Center

Past edition: ICCV 2025

Speakers

Dr. Guodong Ding (National University of Singapore)

Call for Papers

We invite contributions related to, but not limited to, the following areas:

Algorithms & Applications for Skilled Human Activity and Performance Assessment
- Human Skills/Performance/Action Quality Assessment/Evaluation in various domains such as Sports, Industry/Assembly line, Musical and Dance performance, Surgical procedures, Fitness, Robotics, Gait Assessment, Rehabilitation, Healthcare etc.
- Action and Process Mistake Detection using video and video analytics
- Missing Action Prediction and Alternative Action Sequence Prediction.
- Brave new ideas and novel applications of Action Quality Assessment in: Video Generation, Human-Robot-Collaboration, Robot Skills Learning etc.
Feedback Report Generation for Skills/Performance Improvement & Guidance
- Captioning and Narration for Natural Language Corrective Feedback Generation
- Visual/AR/VR-based Corrective Feedback Generation
New Datasets & Benchmarks
- Datasets and Benchmarks
- Multimodal, Multiview datasets
- Creation of Synthetic Fine-grained Action datasets
- Fine-grained Video/Action generation
Metric learning for Skills/Action/Motion Quality Assessment
Fine-grained Action Recognition & Segmentation
Representation Learning & Learning Paradigms
- Pose Estimation and Human-Centric Representation Learning, especially, suitable for challenging scenarios such as high-speed, non-regular actions, occlusions, etc.
- Generalization/Transfer Learning/Domain Adaptation for SAUAFG
- Semi/Self-Supervised Learning for SAUAFG
- Continual/Incremental Skills and Action Quality Assessment
- Incorporation of Knowledge Graphs
- Learnable Task Graphs
Explainability & Interpretability in SAUAFG models
Ethics & Fairness in SAUAFG
- Ethical Considerations, Fairness, and Bias in SAUAFG systems, and strategies to mitigate them.
- Social Impact of SAUAFG, including its ability to democratize access to training and Performance Evaluations, particularly in underserved communities.

Paper Submission

We have following two Paper Tracks. Please use this Openreview Link to submit your papers, and select the appropriate track in OpenReview. We recommend you to use Institution emails to register your account on Openreview, as non-institution emails can take upto 2 weeks to register.

Decisions: March 31, 2026 (All times are in Anywhere on Earth (AoE) timezone)

Notifications sent to authors: March 31, 2026

Cameraready deadline: April 9, 2026. Instructions have been emailed to authors.

1) Proceedings Track. Papers accepted to this track will be published in the official proceedings. Please use CVPR 2026 paper templates. Page limit, excluding references, is 8 pages. Please strictly follow CVPR 2026 paper guidelines. Supplementary material is allowed. We will follow single-stage review process--we will not have rebuttal stage. Tentative Submission Deadline is: March 21, 2026. Camera-ready dealine: April 09,2026.

2) Non-Proceedings Track. These papers would be presented at the workshop, but not included in the proceedings and can also be submitted to future venues. We also invite papers previously accepted to CVPR 2026 or any other venues. This tracks allows to share your work and receive feedback on your work, making it stronger. Work-in-Progress or Late breaking work providing provocations for new work and ideas to emerge, with or without an accompanying evaluation, can also be submitted. Submissions to Non-proceedings Track are not required to be anonymized. Supplementary material is allowed. We will follow single-stage review process--we will not have rebuttal stage. Tentative Submission Deadline is: March 21, 2026. ~~We may consider opening Round 2 for Non-Proceedings Track after Round 1 review. More details to follow.~~ We have received a strong number of submissions during the initial round, so we won’t be holding an official second round of submissions. However, if you are interested in presenting your work, please feel free to reach out to us directly, and we will be happy to discuss potential opportunities.

Please let us know if you face any problems or have any questions.

Workshop Program

Time	Program
13:00 – 13:15	Opening Remarks
13:15 – 13:45	Keynote-1: Dr. Walterio Mayol-Cuevas — Topic: Skill Evolution
13:45 – 14:05	Keynote-2: Dr. Anwesa Choudhuri — Topic: MedGRPO: Multi-Task Reinforcement Learning for Heterogeneous Medical Video Understanding
14:05 – 14:35	Group 1 Oral Presentations + grouped QA
14:35 – 15:05	Group 2 Oral Presentations + grouped QA
15:05 – 15:20	Coffee Break
15:20 – 15:50	Keynote-3: Dr. Guodong Ding — Topic: From Action Segmentation to Skill Understanding
15:50 – 16:20	Group 3 Oral Presentations + grouped QA
16:20 – 16:50	Group 4 Oral Presentations + grouped QA
16:50 – 16:55	Closing Remarks
16:55 – 18:00	Poster Session in ExHall A (please don't hang posters beforehand, as boards are shared amongst workshops this year. Poster boards 281-292 (front+back) are ours. You may board of your choice, we are not following any particular ordering.)

SAUAFG 2026 Workshop Oral Presentation Grouping

Group 1
1. Behavior-based Skill Assessment for Open Surgery from Multi-view and Egocentric Videos
2. SkillSight: Efficient First-Person Skill Assessment with Gaze
3. Can Vision Language Models Judge Action Quality? An Empirical Evaluation
4. Local Brushstroke Quality Assessment via Vision-Language Feedback
5. Toward Fine-Grained Basketball Action Understanding via Broadcast–Tracking Sensor Fusion
Group 2
1. How to Take a Memorable Picture? Empowering Users with Actionable Feedback
2. Sports Form Anomaly Detection Using Skeleton-PatchCore
3. Learning Coach’s Tactics: Skill-aware Readiness Modeling for Interactive Badminton Training
4. Structured Relational Reasoning for Group Activity Assessment
Group 3
1. ViterbiPlanNet: Injecting Procedural Knowledge via Differentiable Viterbi for Planning in Instructional Videos
2. Combining Boundary Supervision and Segment-Level Regularization for Fine-Grained Action Segmentation
3. Distance-Aware Attention for Surgical Task Classification with Video Transformers
4. EgoIndAssembly: Diagnosing Fine-Grained Temporal Hand Reasoning in Vision–Language Models for Skilled Assembly
5. UDAPose: Unsupervised Domain Adaptation for Low-Light Human Pose Estimation
Group 4
1. A Study of Failure Modes in Two-Stage Human–Object Interaction Detection
2. Efficient and Intrinsically Interpretable Spatiotemporal Transformer with Gated Fusion for Group Activity Recognition
3. DUAL-Pose: Efficient Dual-Branch Graph Networks for Skeleton-Based Yoga Pose Recognition
4. Counterfactual Action Quality Assessment for Explainable Skill Feedback