Deployments

On this page you can browse, manage and annotate all deployments within your project.

Deployments table

Each row in the deployment table represents a deployment and includes the following information:

  • Deployment ID: randomly generated identifier for the deployment.

  • Location: the deployment location.

  • Start: the deployment start timestamp.

  • End: the deployment end timestamp.

  • Sequences: number of sequences in the deployment.

  • Imported on: timestamp when the deployment was imported into Agouti.

  • Imported by: the user who imported the deployment.

  • UTC offset: offset in hours from UTC time.

  • Invalid: indicates whether the deployment has been marked as invalid.

  • Tags: any tags assigned to the deployment.

  • Groups: any groups assigned to the deployment.

  • Assigned to: the user assigned to the deployment.

  • Progress: annotation progress. In blue: sequences that were annotated by project members. In green: sequences that were annotated by the AI.

Use the three dots on the right of the column headers to choose which columns are displayed in the table.

Deployments can be filtered and sorted using the controls in the top row of the table. If there are multiple pages of results, use the pagination controls below the table to navigate between pages.

On the right side of each deployment row, click the Edit button to modify any information that was provided when the deployment was created. You can also delete the deployment by selecting Delete at the bottom of the edit deployment page.

Next to the Edit button is the View button. This opens an overview of the deployment, including all deployment details and the sequences it contains.

Manual annotation

To annotate a deployment by hand, click the Annotate button on the right-hand side of a deployment row. This opens the annotation page.

If an AI model for automatic classification has been set in the project settings, the button annotate by AI is also available. Clicking this sends the deployment to the AI queue, which will process the deployment and add observations in the background. You cannot work on a deployment manually when it is in the AI queue.

The annotation page is divided into several areas, each with its own role.

Image viewer

The current sequence is displayed in the centre of the page.

If a sequence contains multiple images, you can browse through them using the arrow keys or by clicking on the arrows on the sides of the image. Images can be zoomed using the mouse wheel and panned by dragging.

Above the image are controls to:

  • Adjust brightness and contrast.

  • Add or remove the current image from your favourites.

  • Reset the zoom level.

  • Reset brightness and contrast settings.

  • Download the current image.

  • View the metadata of the current image

Observation types

When viewing an unannotated sequence, shortcut buttons for Animal, Blank, Setup/Pickup, Vehicle, Deployment Calibration, and Unknown are available.

  • Animal: one or more animals are visible in the sequence.

  • Blank: the sequence does not contain any animals or other relevant content.

  • Setup/Pickup: the camera is being deployed, serviced, or removed.

  • Vehicle: a vehicle is present in the sequence.

  • Deployment calibration: used for density estimation. See the dedicated page that describes everything that is needed for this to work.

  • Unknown: something is present in the sequence, but it cannot be identified with sufficient confidence.

Selecting one of these options annotates the sequence accordingly.

Animal observations

Selecting Animal opens an observation form where you can record information about the species present. Mandatory fields include the species and the amount. Optionally, you can add information about the sex and age of the animal. You can also indicate that the observation is uncertain. Finally, there is a notes section. Any behaviours added in the project settings will also be available in this form.

Animal observations that have already been added to the sequence are shown in the top-right corner of the page. Use Add species to add additional observations to the sequence. Existing observations can be edited or deleted using the buttons next to each observation.

AI-generated animal observations are marked with a small robot icon. The AI model used to make the observation and its confidence score are displayed below the observation.

Make sure to follow the project’s protocol when adding observations. Contact your project coordinator if you are not sure what information is required and optional. In general, we recommend entering at least the species and amount.

If you have the amount of species shortcuts in your personal settings set to something other than zero, shortcut buttons for recently entered species will appear as you work through the sequences. Use these to quickly add an observation of that species.

Automatic annotation

Agouti offers several options to automatically annotate your sequences and add observations. The AI functions are available through Agouti’s user interface, at the click of a button.

In Agouti multiple AI models are available to choose from and most are aimed at species classification. The base for each model is a simple classifier that checks if images are blank, contains an animal, or contains a human. If an animal is detected, a second classifier attempts to identify the species in each image. If a sequence contains multiple images these image-level detections are then further processed to yield a sequence-level observation. In this last crucial step, we apply some decision rules to improve the quality of the AI observations.

The result of this classification process is saved in Agouti as a regular observation. The only difference is that it was added by the AI instead of a user, and it’s marked as such. This makes checking the AI observations for errors easier. You don’t need to check all AI observations, but can use a filter to make a selection. For example, you may only be interested in carnivores so check all of those observations, but simply accept all AI observations of ungulates without checks. It’s up to you to decide on a strategy.

As with any AI there are some things to keep in mind. AI classification quality varies and we recommend to at least manually check a selection of AI observations in your project. There is great variation between projects and differences such as image quality, quantity, context, vegetation etc, all of which can affect AI performance. Furthermore, AI in itself is rapidly advancing which makes it challenging to keep up with the latest & greatest. We hope to find a balance between AI performance, quality and usability for users.

AI functionality in Agouti is in active development and not perfect. Still, we hope that it’s an useful tool and speeds up the annotation process. We recommend that you use the latest version available for a model whenever possible. Any feedback on the performance of a specific model is welcome. Also, if you have any data to share that we can use to improve a certain model that is highly appreciated.

If you have built your own classification model and would like to use or offer that in Agouti that is possible too. Contact us at agouti@wur.nl to discuss.

How to use the AI?

There are two ways to use the AI functionality in Agouti. The first one works at the deployment level and processes all images in the background. This is ideal if you have multiple deployments and don’t need immediate results. The second option is the realtime AI which is available during manual annotation.

To use the deployment-level AI:

  1. In your project settings, scroll down to automatic annotation and select a model from the list. Details about each model can be found below. This list may be expanded over time, and we regularly release new version of models.

  2. Go to the Deployment menu. For each row there is now an additional button: annotate by AI. If you click this button, the deployment is sent to the queue for automatic processing. After clicking, please wait until the text that indicates your position in the queue appears. This can take up to 2-3 minutes for large deployments. Deployments cannot be manually annotated while the AI is working.

  3. When the AI is done, you will see the progress bar is partly green and the status has been updated to ‘AI is done’. You can now manually annotate the remainder of the deployment. You can also use the Observations page to browse through the AI observations.

To use the realtime AI:

  1. In your project settings, scroll down to automatic annotation and select a model from the list.

  2. Open a deployment for manual annotation.

  3. With a sequence open, click the button Analyse on the right-hand side of the screen. This sends the sequence to the AI. Depending on the amount of images in the sequence the AI will suggest one or more observations. You can accept, remove or edit the observations as needed. It can take up to 30 seconds to produce output. If the loading bar completes and there are no observations suggested, the AI was not able to produce any output with enough confidence.

Because AI processing requires substantial computing resources, Agouti uses a queueing system. Depending on the number of deployments waiting to be processed, your deployment may be annotated anytime between several hours to several days. To avoid delays, we recommend submitting deployments for AI processing as soon as possible. If your deployment has not been processed after 72 hours, you can contact us to check its status. The realtime AI is still experimental and may not always provide immediate results. We are continuously working to improve both the speed and quality of our AI models.

Available AI Species models

While we develop AI models ourselves, we try to include works of others as well. There are many high quality initiatives around, some of which have kindly shared their models with us so we can offer them in Agouti. Have a model of your own that you would like to use or contribute? Contact us at agouti@wur.nl.

Generic human/blank model (MegaDetector) v5a + v6

Developed by: Microsoft (until 28 April 2023) Model that supports four classes: animal, human, vehicle and blank. Good option when a specific model is not yet available for your locations. MegaDetector documentation

Agouti Blank/Human/Vehicle v1

Developed by: Agouti.eu
Model that supports four classes: animal, human, vehicle and blank.

DeepFaune (France) v1.2 + v1.4

Developed by: Institut National d’Ecologie et d’Environnement (INEE) of the CNRS
Model that covers 60 species common to France. Highly recommended for other parts of Western Europe as well. DeepFaune documentation

Europe v5

Developed by: Agouti.eu
Model covering 100+ species of mammal and bird that occur in Europe. The model has been trained with data from across all corners of Europe, with a much wider range and quantity of images than the model for Western Europe. It also leverages a new underlying architecture that works better when there are multiple animals in view. Because this model is a bit more broad than the Western Europe model it may or may not perform better on your data. We recommend trying out both, depending on where you are working.

Europe v6

Developed by: Agouti.eu
This model has been trained with data from Western Europe, mostly the Netherlands, Belgium, Germany and Luxembourg. It covers about 139 taxonomic units. Some species were grouped to a higher level in the taxonomy because they cannot be reliably told apart by the model at this point. For example, Martes or Turdidae. See species list.

French Guiana v1

Developed by: Agouti.eu
Model that covers 10 species that are common in French Guiana. See species list.

India v2

Developed by: Agouti.eu
Model that covers 11 species common to India. See species list.

Nepal (vertical) v1

Developed by: Agouti.eu
Model that covers 13 species common to the lowlands of Nepal (Terai). This model is designed for use with camera that are placed facing down vertically.

Panama v2

Developed by: Agouti.eu
Model that covers 17 species common to central Panama. See species list.

Eastern Africa v1

Developed by: Agouti.eu
Model that covers 50 species that occur in Eastern Africa. See species list.

Southern Africa v3

Developed by: Agouti.eu
Model that covers 32 species that occur in Southern Africa. See species list.

New Zealand v1

Developed by: Agouti.eu
Model that covers 24 species that occur in New Zealand. See species list.

SpeciesNet v4.0

Developed by: Google / Wildlife Insights team
This ensemble model combines MegaDetector for object detection with the SpeciesNet classifier trained on over 65 million camera trap images from diverse geographic regions. It covers 2000+ labels including animal species, higher-level taxa (e.g., Mammalia, Felidae), and non-animal classes (blank, vehicle). For more information, see Wildlife Insights or the SpeciesNet repository.

Voedertonnen v1

Developed by: Agouti.eu
This model was trained with images from Belgium. It identifies common species at feeders and also detects the ‘voederton’ (feeder) itself. See species list.

Keyboard shortcuts

Agouti has some basic keyboard shortcuts that can be used to annotate more quickly:

During annotation, you can navigate by using:

  • N for Next: move to the next sequence

  • U for Next unannotated

Further controls:

  • Left and right arrow keys select the previous/next frame of the sequence

  • By default the Animal button is selected, by using tab or shift + tab you can move to the next / previous button

  • On pressing Enter the button is activated, e.g. the Animal form is opened or the sequence is marked Blank/Empty

  • On Escape the form is closed (or canceled if not yet saved)

When opening the Animal form:

  • If no species is selected, the Species selector is focused

  • If a species is selected, the Amount box is selected

  • Use tab / shift + tab to move through fields

  • Dropdowns are expanded when in focus

  • Species shortcuts can be used by using keyboard keys 1, 2, 3, etc.

When opening an already annotated sequence:

  • The Valid box is selected by default

  • Use tab / shift + tab to select an observation and press Enter to open it