ICIP 2025 Workshop on Computer Vision for Ecological and Biodiversity Monitoring (CM-EBM)
We are facing a global environmental crisis caused by anthropogenic climate change, and the degradation of habitats and ecosystems due certain agricultural practices and urbanization. To understand the extent of this impact and environmental responses to intervention, it is necessary to monitor ecosystems through the collection and identification of patterns in various data, including remote sensing imagery, ground-level imagery, and video.
This workshop will bring together leading experts from both the computational and ecological research communities to share the latest innovations, datasets, and applications in a focused forum with two keynote speakers alongside regular paper submissions.
The CV-EBM workshop will be a full-day event, and all submitted papers undergoing the standard review process. Accepted papers will be included in the ICIP proceedings.
Call for Papers
We are now accepting regular paper submissions to CV-EBM!
Topics include (but are not limited to):
Scene analysis
- Fine-grained/hierarchical classification
- Habitat type/biodiversity classification
- Land-use/land-cover mapping
- Biomass/carbon sequestration estimation
Object detection & segmentation
- Detection & counting of plants/animals
- Handling small objects and occlusions
- Segmentation of camouflaged organisms
- Organ/anatomical segmentation
Technical challenges
- Data-efficient learning
- Learning from noisy labels
- Multi-modal learning
- Continual/lifelong learning
Imaging modalities & benchmark datasets
- Ground-level imagery
- Remote sensing/earth observation data
- Camera trap/video data
- Hyperspectral/multispectral imaging
- Microscopic imaging
Tracking and temporal monitoring
- Animal re-identification
- Multi-object tracking (MOT)
- Tracking in challenging environments
- Population monitoring
- Pose estimation
Applications
- Species monitoring and identification
- Predictive modelling of habitat loss/change
- Prioritising areas for conservation/restoration
- Monitoring harmful/invasive species
Important dates
Satellite Workshop Paper Submission Deadline: 28 May 2025
Satellite Workshop Paper Acceptance Notification: 25 June 2025
Satellite Workshop Final Paper Submission Deadline: 2 July 2025
Satellite Workshop Author Registration Deadline: 16 July 2025
Main Conference Dates: 14 – 17 September 2025
CV-EBM Satellite Workshop Date: TBC
Organising committee
University of Lincoln, UK
- Dr James Brown, Associate Professor in Computer Science (He/Him)
- Dr Petra Bosilj, Assistant Professor in Computer Science (She/Her)
- Dr Wenting Duan, Assistant Professor in Computer Science
- Dr Lan Qie, Assistant Professor in Ecology and Conservation
- Dr Hongrui Shi, Postdoctoral Research Associate
- Mx Villanelle O’Reilly, PhD Candidate (They/Them)
University of Oxford, UK
- Dr Katrina J Davis, Associate Professor in Conservation Biology
- Dr Rob Salguero-Gómez, Associate Professor in Ecology
- Prof Ben Sheldon, Professor of Ornithology
- Prof Graham Taylor, Professor of Mathematical Biology
- Dr Georgios Voulgaris, Postdoctoral Researcher in Deep Learning and Spatial Ecology
Our statement on diversity
We welcome submissions from authors of all backgrounds regardless of race, ethnicity, gender, sexual orientation, disability, and geographic location.
We especially encourage authors from underrepresented groups (minority ethnicities, women, LGBTQ+) to attend and contribute.
Extended Abstract
We are facing a global environmental crisis caused by anthropogenic climate change, and the destruction of habitats and ecosystems due to agriculture and urbanization. To understand the extent of this impact and environmental responses to intervention, it is necessary to monitor ecosystems through the collection and identification of patterns in various data, including remote sensing imagery, ground-level imagery, and video. State-of-the-art computer vision based on deep learning is an emerging technology for ecological applications and offers the potential to automate the analysis of many different species, habitats, their relationships, and their responses to intervention and management.
Computer vision systems have been successfully developed for ecological monitoring, however there are various open challenges that limit their applicability in real-world scenarios at scale; (i) handling geographic differences in species composition (domain adaptation/generalisation), (ii) lack of high-fidelity annotations (weak supervision), (iii) utilising and fusing information from different data sources (multi-modal learning), (iv) leveraging large amounts of unannotated data (self-supervised learning), (v) creation of high-fidelity publicly available benchmark datasets (e.g., Pl@ntNet).
Given the volumes of image and video data now routinely collected and catalogued by ecologists, there is significant untapped potential to develop approaches that can perform national or even global scale analyses that would otherwise be intractable through manual methods. This workshop will bring together leading experts from both the computational and ecological research communities to share the latest innovations, datasets, and applications in a focused forum with two keynote speakers alongside regular paper submissions. This interdisciplinary workshop is aligned with the UN’s Sustainable Development Goal 15, “Life on Land”, which aims to “Protect, restore and promote sustainable use of terrestrial ecosystems, sustainably manage forests, combat desertification, and halt and reverse land degradation and halt biodiversity loss.”