Webinar: Using computer vision to keep track of animals in the wild

On Thursday 29 October our Solution Partner Vantage AI organizes the webinar ‘Using computer vision to keep track of animals in the wild’ together with Sensing Clues.

Our Wildlife Insights tool suite combines data from various sources to offer wildlife rangers a detailed view of he current state of affairs in their wildlife reserve.

As with any other data-driven tool, ‘the more quality data the better’-rule applies. Traditionally, much relies on observations by rangers in the field. Unfortunately, however, boots on the ground are always scarce in comparison with the vast areas that need protection.

A common additional source of information is the use of camera traps that capture wildlife on film. However, extracting valuable information from this footage often turns out to be a quite labor-intensive process as ecologists have to manually sift through hours of camera trap footage.

In this webinar, Mike Kraus of Vantage AI and Jan Kees Schakel of Sensing Clues will show how Vantage AI developed an image recognition solution for Sensing Clues that resolves this issue by extracting meaningful insights from camera trap footage in an automated, resource-efficient manner.


Details

When: Thursday 29 October - 10AM CET
Duration: 1 hour
Presentation by: Mike Kraus (Vantage AI) and Jan-Kees Schakel (Sensing Clues)


What wel'll cover

  • Introduce the Sensing Clues wildlife insights platform and show how rangers use it on a day-to-day basis.

  • Touch upon the high-level workings of image recognition algorithms.

  • Zoom into the process of developing a solution that addresses the two distinctive steps of detecting and classifying animals in the wild.

  • Elaborate on the challenges that come with each step. For example, dealing with motion blur and bad illumination.

  • Illustrate how we are working to set up a human-in-the-loop framework that will allow the algorithm to learn from human feedback on its predictions.


For whom

This webinar will be of interest to data scientists who would like to know more about image recognition and its potential, whilst simultaneously getting an understanding of the aspects that come into play when putting this technique to action.