1 #TropiCon20 #CameraTraps are a popular wildlife monitoring tool that take photos or videos when triggered by an animal @MammalWeb is a citizen science project that invites the public to upload and classify camera trap footage, but is it easier to classify photos or videos?

2 #TropiCon20 15 pairs of camera traps were placed at 74 sites in the Forest of Dean over 18 weeks, with one camera in each pair taking photos and the other videos, switching settings halfway through each deployment. Footage was uploaded to separate photo & video projects.

3 #TropiCon20 Accuracy of MammalWeb classifications were determined for photo sequences and videos of different lengths. The video project has received a higher percentage of correct classifications than the photos, with 5 second videos receiving the highest % of correct answers

4 #TropiCon20 Changes in accuracy between photo and video footage varied with species with some, e.g. grey squirrel, seeming to be easier to spot in videos. Others e.g. fallow deer, were correctly identified in a slightly higher percent of classifications in sequences of 8 photos

5 #TropiCon20 Citizen science projects can help engage the public and provide valuable ecological data. Using video instead of photo in camera trap citizen science projects could increase classification accuracy thus improving data quality and leading to more effective monitoring

P.S. massive thanks to @gloswildlife, in particular @JosieBridges91 and @mcnicol27 for their help , and of course everyone who has classified footage on @MammalWeb so far. We're still collecting data and will be exploring it further in the future - so stay tuned for updates!