AI initiative registry

71 current initiative(s)

• 151

  Zooplankton image analysis

  Use Random Forest and residual neural networks (ResNet) for automated classification of zooplankton images collected using: 1) scanning of preserved samples (ZooScan); and 2) in situ imagery (Underwater Vision Profiler, UVP). Imagery focused on surveys and samples collected along the West Coast of Vancouver Island and offshore from 2022-present.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 104

  Zooplankton classification from Video Plankton Recorder

  Using machine learning algorithms to identify plankton for rapid zooplankton classification from the Video Plankton Recorder, in support of efforts to assess North Atlantic Right Whale foraging habitat.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 23

  Video Summarization for Aerial Surveillance

  AI pilot under development to automatically detect and flag relevant events (e.g., human/vessel activity) in surveillance video for information management. Action plan for this initiative: 1. Governance: Oversight by a multidisciplinary group, routine updates and risk assessments to DM. 2. Guardrails: Pilot with departmental data only, audit outputs for accuracy and bias. 3. Work processes: Deliver targette

  Pilot Programs Sector Perception and Understanding

• 96

  Use of machine-learning (ML) algorithms and high-performance computing (HPC) for the optimization of satellite image analysis for mapping coastal habitats

  This project aims to improve satellite image analysis of marine habitats leveraging NRC HPC and ML assets to apply cutting-edge ML techniques to the challenges of remote sensing data analysis. Relates to ongoing project (below) through MPC service-level agreement.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 118

  Use of artificial intelligence models to facilitate seafloor video annotations in conservation areas.

  The purpose of this initiative is to utilize AI to facilitate seafloor video annotations collected as part of the MCT and Benthic Ecology programs in the NL Region. A large number of vidEcosystems and Oceans Science is collected annually, and manual video annotation (e.g., locating and counting observations) is extremely time-consuming. A published “object detection model” (FathomNet Megalodon Detector, YOLOv8x) trained using marine taxa by MBARI is currently being used to identify objects in our seafloor images. The model does not identify the objects (e.g., to species), but it largely accelerates the process given that locating objects is the most time-consuming part of the work.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 81

  Topaz AI Software

  Enhance underwater video from ROV and tow camera deployments in the offshore region and elsewhere. The deployment of these platforms is incredibly time- and resource-consuming, making the optimization of available video an important part of getting the most from this imagery. In addition there are camera limitations from the available DFO survey tools and framegrabs from video will be used as a major product from surveys, making enhancement to high quality images crucial. Topaz Video AI and Topaz Photo AI

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 240

  The Current

  Don't let your harvested data get stuck in still water. Navigate new data streams with AI automation, augmentation and information enablement. The Current is a cloud-hosted platform that enables users anywhere in the department to ingest large volumes of structured and unstructured documents, extract data using optical character recognition and large language models, validate and review results through human-in-the-loop workflows, and relate extracted information using shared concepts and standards. The architecture emphasizes asynchronous processing, modular extraction services, strong traceability, and governance-ready design. The platform supports configurable extraction tasks, versioned models and definitions, auditable review processes, and future discovery of related research outputs and datasets across heterogeneous sources. As the initial use-case, the Text Intelligence and Data Extraction (TIDE) Service has been developed for the Salmon Habitat Restoration (SHARE) System. TIDE is a baseline iteration of The Current that fulfills the minimum viable product requirements of SHARE. TIDE focuses on digitizing, extracting, and modeling data from large unstructured text documents. TIDE establishes the baseline architecture and demonstrates a scalable workflow in Production.

  Development Fisheries Management Perception and Understanding

• 68

  Stereo camera image analysis

  Tools built in python to automate image analysis and fish measurement tasks. Applications include a conveyor belt camera and underwater towed stereo camera system Computer vision tools (Python, Viame)

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 189

  Site Assessments

  Automatically assess presence of eel grass from drone video Solution being developped to assist Fish and Fish Habitat Protection Program in reviewing hundreds of hours of drone footage to detect presence of Eel Grass.

  Ideation On hold Indigenous Affairs, Aquaculture and Governance Perception and Understanding

• 228

  SILScanner

  Deploy on DFO's PBMM Cloud a proof of concept application that uses optical character recognition (OCR) to digitize and validate Stream Inspection Logs (SILs). The PoC was originally developed on SSC's cloud infrastructure. It is now being deployed in DFO PB cloud infrastructure. BLADES initiative and IT security teams are involved.

  Pilot Chief Digital Officer Perception and Understanding

• 97

  Satellite image analysis for mapping eelgrass habitat in the southern Gulf of St. Lawrence.

  This project supports Gulf Region needs for spatial data representing eelgrass habitat through a service level agreement with Marine Planning and Conservation. Machine learning algorithms have been applied for classification of satellite imagery. MPC and Science have been collaborating for several years, annually purchasing satellite imagery, with classification conducted by both groups.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 132

  Salmon Scale Aging​

  A model using computer vision to look at images of salmon scale and predict the age of salmon and also explain how it arrived at its prediction.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 49

  Salmon Fence Counting​

  A model using computer vision to analyze water stream fence video-footage, to simultanously count the number of and predict the species of salmon.​ YOLO 11 Chinook trained on 735 annotated frames from 63 videos Coho trained on 432 annotated frames from 44 videos Sockeye trained on 1149 annotated frames from 36 videos Chum untrained (annotation of 1368 frames from 43 videos in progress) Pink Humpback

  Pilot Programs Sector Perception and Understanding

• 50

  Salmon Data Digitization​ (See B22, B26)

  Digitizing non-machine-readable (handwritten) documents and extracts characters into a digital format so that it is accessible, editable, and useable for analysis. Document Intelligence - Custom Classification and Extraction Model Azure OpenAI GPT4o mini Data and Information from ~45,000 pages have been digitized

  Pilot Programs Sector Perception and Understanding

• 130

  Salmon Data Digitization​

  Digitizing non-machine-readable (handwritten) documents and extracts characters into a digital format so that it is accessible, editable, and useable for analysis.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 161

  Remote sensing of the internal tide in the St. Lawrence

  Using model outputs (STLE500/STLE200) to 1) help identify internal tide surface signature in SWOT (Surface Water and Ocean Topography) data and 2) train an AI to detect internal tide in SWOT data.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 224

  Remote Observation of Watercraft on Roads

  The Aquatic Invasive Species Program is proposing to launch a pilot initiative aimed at addressing a critical knowledge gap in our understanding the spread of aquatic invasive species. Invasive species can attach to boats, trailers, and related equipment—a process commonly referred to as the "stowaway pathway"—which enables their unintentional spread across ecosystems when watercraft are transported between water bodies. The AIS Program is seeking to understand watercraft movement into Canada via international land border crossings. Currently, the extent of this "stowaway" pathway remains largely unquantified. To address this, the pilot will deploy remote camera systems at select border entry points to capture imagery of incoming vehicles. Leveraging machine learning, the system will automatically identify and flag images containing watercraft. This data-driven approach will provide actionable insights to inform strategic decisions on the timing and placement of watercraft inspection and decontamination resources. This initiative may represent a significant step toward enhancing our ability to proactively manage aquatic invasive species risks at key entry points.

  Pilot Perception and Understanding

• 58

  Phytoplankton image classification model

  Create a neural network model to classify phytoplankton images collected by in-situ phytoplankton imaging sensors. There are two discrete steps to this project: 1) create a library of annotated images for model training; 2) creation and refinement of the CNN model. Convolutional Neural Network, InceptionV3 and ResNet

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 74

  Pacific Plankton Monitoring: Imagery and automated classification

  • Monitoring of zooplankton and phytoplankton in the water column with shipboard and benchtop plankton imaging instruments (UVP, ZooScan, PlanktoScope). • Application of automated classification of survey and preserved sample image data sets to complement and enhance spatial and temporal resolution of monitoring in Canada’s EEZ and offshore NE Pacific. Development of and regular optimization of automated image classification models using AI: deep learning (residual networks) and machine learning (random forests) methods.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 95

  Optimization of Machine Learning modelling strategy for coastal zooplankton taxonomic identification from flow microscopy imaging.

  As part of Aquaculture Monitoring Program we are using Ecotaxa's ML algorithm to identify zooplankton. The project aims at identifying the best strategy in building model training sets for each region monitored. This strategy will be the basis for future AMP zooplankton identification work.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 66

  Optical Character Recognition Data transcription tool (See B6, B26)

  Built a proof of concept tool to facilitate data transcription and validation from hand-written field notes. Tested on ~100 year old salmon observations and present-day salmon stream inspection logs to determin ethe cost/feasibility/effectiveness of implementing such a tool in Salmon monitoring contexts. (Digitizing non-machine-readable (handwritten) documents and extracts characters into a digital format) Computer vision (azure document intelligence),

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 147

  Optical Character Recognition Data transcription tool

  This completed project built a proof of concept tool to facilitate data transcription and validation from hand-written field notes. tested on ~100 year old salmon observations and present-day salmon stream inspection logs to determine the cost/feasibility/effectiveness of implementing such a tool in Salmon monitoring contexts.

  Pilot Ecosystems and Oceans Science Perception and Understanding

• 245

  Object, Classification, Tagging, Organization, and Pattern Understanding Solution (OCTOPUS)

  This initiative leverages AI technologies to build an Object, Classification, Tagging, Organization, and Pattern Understanding Solution (OCTOPUS) to enhance the analysis of video, image, and audio data, improve monitoring capabilities and support decision-making processes at DFO. The MVP will initially focus on video processing capabilities, with planned expansion to image and audio analysis in FY 2026–27. In addition, the initiative will deliver integrated data annotation and model training tools to support continuous improvement and reuse of AI models.

  Development Chief Digital Officer Perception and Understanding

• 33

  Machine-learning-assisted Benthic Marine Life Annotation Tool

  The goal of the project is to improve the efficiency of the image annotation process to reduce manual efforts of the benthic science team and reduce time-to-insight from underwater surveys. Further development is prevented by DFO's IT infrastructure. There are ongoing discussions with CDOS to resolved this but the project is on hold in the meantime. Client: Quebec Region/Pacific Region - Oceans and Ecosystem Sciences Division

  Ideation On hold Strategic Policy Perception and Understanding

• 91

  Localization and Tracking of Beluga Whales in Aerial video Using Deep Learning

  Deep-learning model was developed us YOLOv7 to detect beluga whales from imagery footage. The main contribution of this research is providing a system that accurately detects and tracks features of beluga whales, both adults and calves, from aerial footage.

  Pilot Ecosystems and Oceans Science Perception and Understanding