The rapid evolution of Large Foundation Models (LFMs) has transformed the landscape of biomedical research, clinical decision-making, and healthcare innovation. From decoding complex biological interactions to assisting in diagnosis and drug discovery, LFMs have demonstrated remarkable potential across a broad spectrum of biomedical applications. However, their adaptation to this highly specialized and sensitive domain presents unique challenges ranging from data scarcity and heterogeneity to issues of interpretability, fairness, and reproducibility. The Workshop on Large Foundation Models for Biology and Biomedicine (LFMBio 2026) aims to bring together researchers, practitioners, and industry experts to advance the science and practice of applying LFMs to biomedical problems. The workshop will serve as a forum for presenting original research, fostering interdisciplinary dialogue, and exploring cutting-edge innovations in model design, multimodal integration, trustworthiness, and ethical deployment. We invite contributions that span foundational model development, performance optimization, knowledge representation, real-world clinical applications, and the societal impact of these powerful technologies.

Computer vision methods trained on public databases demonstrate performance drift when deployed for real-world surveillance, compared to their initial results on the test set of those employed databases.In this workshop, we are interested in papers reporting their experimental results on any application of computer vision in real-world surveillance and object security, including the protection of buildings and facilities within critical infrastructure, challenges they have faced, and their mitigation strategy on topics like, but not limited to:- Object detection- Tracking- Action recognition- Scene understanding- Super-resolution- Multi-modal surveillanceFurthermore, the workshop has a special attention to legal and ethical issues of computer vision applications in real-world scenarios.We therefore also welcome papers describing their methodology and experimental results on legal matters (like GDPR, AI Act, and US Executive Order on AI) or ethical concerns (like detecting bias towards gender, race, or other characteristics and mitigating strategies).We particularly encourage submissions addressing safety, reliability, and regulatory compliance for critical infrastructure protection, as well as privacy-preserving approaches in high-security environments.The workshop also hosts a competition on robust thermal-image object detection.We have run this workshop previously four times at WACV (2022-2025) and once at ECCV (2022).

The rise of generative AI and foundation models presents new challenges for ensuring robustness against synthetic and adversarial media. Research in adversarial machine learning has shown that detection systems can be bypassed with subtle perturbations, enabling malicious content that undermines societal trust and national security. This workshop offers a venue for advancing work at the intersection of synthetic media forensics and adversarial robustness, with a focus on provenance analysis, fingerprinting, authenticity verification, and resilience across diverse generative architectures. Expected outcomes include a taxonomy of joint synthetic–adversarial threats, benchmark resources for evaluation, and stronger collaboration between technical, forensics, and policy communities.

Clinical medicine represents one of the most demanding testbeds for computer vision, where methods must function reliably under distribution shifts, strict regulatory constraints, and high social impact. Despite rapid progress in foundation models, multimodal learning, and self-supervision, a persistent gap remains between state-of-the-art CV research and its safe, effective deployment in clinical practice. This workshop, 'Pixels to Patients: Bridging CV State-of-the-Art with Clinical Impact', will bring together leading researchers, clinicians, and industry partners to address this gap head-on. Core themes include adapting recent CV breakthroughs -foundation and vision-language models, domain adaptation, and data-efficient learning - to healthcare, and exporting lessons from clinical deployment - bias auditing, monitoring, and regulatory compliance - back to the broader CV community. The program will feature keynotes from pioneers in medical AI, oral and poster sessions, a Problem-Pitch track to seed benchmark challenges, and a panel on real-world deployment and interoperability. By positioning medicine not as a silo but as a proving ground for trustworthy, generalizable vision systems, this workshop aims to catalyze advances that resonate across safety-critical domains, from healthcare to robotics and beyond.

The workshop invites paper submissions focusing on the analysis and interpretation of images and videos captured during winter sports and related summer activities such as mountain sports (e.g., mountaineering, downhill biking, climbing). Topics include video understanding, pose and performance analysis, injury prevention, trajectory and scene reconstruction, crowd monitoring, AR/VR for fan engagement, and dataset creation. We also welcome work addressing challenges such as harsh weather, real-time processing, multimodal fusion, and camera pose estimation in broadcast videos.

The rapid growth of computer vision (CV) and artificial intelligence (AI) is reshaping agriculture, offering new approaches to challenges in food security, climate resilience, and sustainability. Scaling these technologies requires bridging core CV research with domain-specific data infrastructures and practical user needs. Key applications include crop and soil monitoring, pest and disease detection, yield forecasting, and resource optimization. Yet, these tasks are hindered by heterogeneous field data, limited labeled datasets—especially in edge environments—and the demand for models that are both scalable and interpretable.Recognizing these challenges, we launched the HARVEST workshop series (the first hosted at 54th International Conference on Parallel Processing in San Diego, CA, supported by the NSF ICICLE AI Institute led by the Ohio State University) to build community and cyberinfrastructure at the intersection of AI, HPC, and agriculture. The event showcased high-profile keynote speakers, and engaged around 30 participants from academia, national labs, and industry. Our agenda included technical talks, panels, and hands-on demonstrations. Through a partnership with NSF AI Institutes Virtual Organization, we were able to secure travel awards for 12 rising academic researchers, many of whom are now part of a growing expert network community of computer scientists and agricultural experts.Bringing this momentum to WACV, our HARVEST-Vision workshop directly aligns with the community’s focus on impactful real-world CV: agricultural cyberinfrastructure is an urgent, societally relevant application space for advances in robust perception, domain adaptation, scalable datasets, and explainable, deployable AI. By fostering interdisciplinary collaborations, catalyzing new datasets, benchmarks, and reproducible pipelines, and supporting diversity via travel awards, our workshop will expand WACV’s reach, accelerating knowledge transfer and enabling high-impact computer vision research for agriculture and beyond.

The growing global concern around waste management, illegal dumping, and environmental pollution highlights the urgent need for intelligent monitoring solutions. Advances in computer vision, guided by the impressive progress in artificial intelligence technologies, offer promising opportunities to address these challenges. However, the scientific literature in this field points out that key research gaps remain, including the lack of robust detection methods for diverse environments, limited datasets and benchmarks, and the need for solutions that can be deployed in real systems with limited computational resources running in real time.The International Workshop on Smart Waste Monitoring (WasteVision) will provide a unique forum for researchers and companies to present and discuss novel contributions in this emerging field. The workshop seeks to advance the state of the art in smart waste monitoring, illegal dumping detection, and environmental pollution surveillance while fostering interdisciplinary collaboration.We invite original research contributions in (but not limited to) the following areas:- Image analysis for waste detection and classification- Video analysis for waste tracking and management- Computer vision methods for detecting illegal waste disposal- Multimodal systems for dumping identification- Remote sensing and UAV-based waste monitoring- Video and image analytics for pollution tracking- Datasets and benchmarks for waste and pollution monitoring- Applications and case studies (real-world deployments in urban and rural contexts)The workshop will also host the first edition of the Illegal Waste Dumping Detection (IWDD) contest, in which the participants will receive a novel dataset for training their approaches for illegal waste dumping detection: the methods will be evaluated on a private test set, in order to ensure a fair comparison on unpublished videos. We welcome scientific papers describing the methods and results obtained during the contest, including innovative approaches, system designs, and comparative analyses.We will accept the submissions of regular papers with more than 5 pages, whose template must follow the same formatting guidelines required by the main conference. All the accepted papers will be published in the proceedings alongside the main conference. More details will be provided on the workshop website.

Can computer vision and multimodal systems forget?

Machine unlearning is often discussed in the context of privacy – particularly as a response to data removal requests in relation to Europe’s General Data Protection Regulation’s right to be forgotten. However, computer vision and computer science technologists rarely have the opportunity to engage directly with privacy and AI governance policy experts in Machine Unlearning (MU) discus- sions. This tutorial changes that! By bringing together researchers with MU technology expertise and others with privacy and AI governance policy expertise the tutorial aims to improve understanding between both groups. Expect presentations of cutting-edge MU approaches and active Q&A and discussion periods including about policy implications and limitations. Invited speakers include ‘YZ’ Yezhou Yang, associate professor, School of Computing and Augmented Intelligence at Arizona State University and Kairan Zhao, PhD candidate and teaching assistant in Machine Learning at the University of Warwick.

The rapid digitization of textual and visual information has made automated document analysis increasinglycritical across industrial, scientific, and cultural domains. Despite advances in computer vision, research largely focuses on limited document types and tasks, leaving challenges such as heterogeneous formats,low-resource languages, non-standard layouts, and historical documents largely unaddressed. This workshop aims to advance document understanding by exploring cutting-edge approaches, including generativemodeling, self-supervised learning, multimodal fusion, and few-shot adaptation. By fostering collaborationbetween computer vision researchers and domain experts, it seeks to promote solutions for generalizationacross diverse document types and low-data scenarios. Currently underrepresented at major computer visionvenues, document analysis will benefit from a dedicated forum connecting the WACV and ICDAR communities. VisionDocs will showcase state-of-the-art methods, stimulate cross-disciplinary exchange, and define new research directions, advancing both scientific understanding and practical AI applications in documentanalysis.

The rapid convergence of event-based vision and generative artificial intelligence (Gen-AI) offers unprecedented opportunities to redefine perception and visual innovation. Event cameras provide asynchronous, high-temporal-resolution data that complements traditional frame-based sensing, while generative models have revolutionized content synthesis, restoration, and reasoning across modalities. The Event-based Vision in the Era of Generative AI (EVGEN 2026) workshop will serve as a forum to explore this synergy, addressing topics such as video generation and interpolation, motion deblurring and prediction, multimodal sensor fusion, gesture reconstruction, and applications in autonomous systems. By fostering dialogue among researchers from neuromorphic vision, computer vision, robotics, and AI, the workshop aims to inspire novel methods, highlight emerging applications, and chart new research directions. EVGEN 2026 will feature invited talks, poster sessions, lightning talks, and a panel discussion, bringing together leading experts and early-career researchers to shape the future of event-driven perception enhanced by generative models.

Motivation and Impact: There is a growing need for venues that foster a richer collaboration between the computer vision and geospatial image analysis communities. By being collocated with WACV (a top-tier computer vision workshop at the cutting edge of computer vision applications), our proposed workshop will provide a platform for algorithm developers and computer vision researchers as well as researchers in the geospatial imaging and image analysis communities to come together and work on challenging problems of great social relevance}. This workshop will build on the foundations of the very successful full-day workshop https://sites.google.com/view/geocv GeoCV @WACV 2025 that was very well received (5 keynote talks, 30 posters, 20 full papers in the proceedings and a full room at the venue). For the 2026 Edition of GeoCV, we propose to emphasize self-supervised learning for the training of large vision and multi-modal foundation models, domain generalization in the context of such models and related concepts as they pertain to Geospatial Image Analysis applications. We believe that this workshop would be a complementary addition to the WACV lineup of workshops by focusing on a niche but very important and emergent area. In addition to the 2025 Edition of GeoCV, we have had a very positive experience hosting a workshop https://sites.google.com/view/morse2025 - (MORSE 2025)} that complemented other workshops and provided synergy in the https://cvpr.thecvf.com/Conferences/2025/workshop-list - {remote sensing track} at CVPR 2025.Expected Outcomes: We expect this workshop to serve as a platform for dissemination of the state-of-the-art in computer vision and AI for geospatial imaging and its applications. Additionally, by bringing together leading researchers from academia and industry to deliver talks on emerging algorithmic ideas, sensing capabilities and research directions, we aim to provide a venue for cross-fertilization of ideas to further the impact computer vision can play in this rapidly growing area of geospatial image analysis, as well as accelerate the adoption of the latest trends and promising developments in computer vision for the analysis of geospatial image analysis at scale. Our workshop represents a diverse team, spanning 5 countries across the globe. Additionally, our team and the planned speakers represent a wide range in terms of their stage-of-career, from rising stars to established researchers. This workshop will also serve as a venue for students and postdocs to network with leading researchers in these emerging and important research areas. Details are provided in the enclosed proposal PDF file, following the WACV workshop proposal template.

Recent advancements in generative models have revolutionized the way researchers approach data-driven tasks. The advent of sophisticated generative models, such as Generative Adversarial Networks, Variational Autoencoders, and Diffusion Models have empowered practitioners to create partially or fully synthetic data closely reflecting real-world scenarios. These generative models' significance lies in their ability to produce remarkably realistic data, thus mitigating challenges associated with data scarcity. As a result, the usage of synthetic data has become increasingly prevalent in various research domains, offering a versatile and ethical alternative for training and testing machine learning algorithms. However, the very realism that makes synthetic data valuable also blurs the line between authentic and manipulated content, resulting in datasets that can potentially be used to mislead, manipulate, or even harm individuals when used unethically. The SynRDinBAS Workshop \& Challenge aims to explore the diverse applications of synthetic realities and data in biometric analysis, while addressing critical security issues such as data privacy and ethical concerns of data manipulation. Participants will examine how synthetic datasets have been instrumental in training systems for facial recognition, emotion detection, gesture recognition, etc. The workshop will showcase exemplary use cases demonstrating how synthetic data not only overcomes limitations of real-world datasets but also fosters the development of more robust and accurate models. Additionally, potential risks and ethical dilemmas that arise from manipulating data will also be discussed, ensuring that our approaches prioritize privacy and integrity in biometric applications. The hosted competition will focus on bridging the research gap associated with the detection of partially synthetic data, as localized changes (such as adding or removing objects or subtly altering faces) are more difficult to detect and more likely to deceive viewers. **Challenge Design:** We will host a script-based challenge, ensuring open access and reproducibility for the research community. At least one of the tasks will focus on detecting of images generated from state-of-the-art models, but the central novelty will be the inclusion of more nuanced manipulation scenarios. Tasks include detecting and localizing object additions/removals, identifying in-painted or altered regions, and distinguishing between fully synthetic, partially synthetic, and pristine content. By anchoring the challenge in more subtle forms of manipulation, we aim to stimulate new methods that move beyond binary classification toward a fine-grained understanding of image authenticity. New datasets will be generated specifically for this effort, with a substantial portion synthetically manipulated at varying levels of granularity, and annotated for both detection and localization. Only small pilot task samples will be released in advance, enabling participants to calibrate their approaches while ensuring that test sets remain robust for evaluation. The models submitted by the participants will be evaluated using testing benchmarks. The organizers will not have access to individual models, ensuring protection of participants' intellectual property. A submission platform will be dedicated to the competition. Submissions will be assessed using a combination of standard detection metrics: accuracy, balanced accuracy, AUC, and localization-specific metrics such as IoU and pixel-level F1 scores. \textbf{Note}: To encourage participation and reward innovation, top performers may be eligible for research grants of up to \$$250,000 and travel stipends will be provided for invited teams.