Mara Animal Detector

Inference pipeline to detect Animals in Mara Images

This page describes how one can run the BisQue Module MaraAnimalDetection

Run Instructions

This module takes in an image or a dataset of images. It generates bounding box predictions indicating the location of several animal classes in the image.

Navigate to Module Page

Login >> Analyze >> Maasai Mara (in Groups Column) >> MaraAnimalDetection

Navigation to Module Page

Expected Inputs

• An image or a dataset of images

  • Click here for a sample dataset of input images

• Machine Learning Model

  • Click here for a sample model file

• Output G-Object Name

  • This can be any string that can be used to name the annotations predicted by this module execution.

Run Inference

Hit Run after providing required inputs.

You should be able to see the status messages at the bottom of the page.

Module page after running the inference

Expected Outputs

• Annotations will be added to the images/dataset as G-Objects

  • Navigate to the images present in the dataset to visualize the predicted G-Object

Output Interpretation

In this section, we describe how to visualize/update the predictions generated by running the model.

Once the module execution is complete, one can navigate to the input dataset to visualize the model predictions in the form of G-Objects.

• Navigate to the dataset.

• Go to the Annotation View.

• Click on Graphical to view the annotation

Export the Predictions

• Navigate to Dataset Page

• Click on Download

• Click on Graphical Annotations as XML