How to use this site
Please note that we are currenty in our beta test phase, and we are updating this site on a regular basis. Your patience with any issues you see is much appreciated and any feedback or questions you have for us is welcome and highly valued!

Navigate to the Observations tab to view images aggregated by date, species, and observer.

Navigate to the Images tab to view all images that have been contributed to the WildTrackAI System.

Images are displayed with associated information received with the image, information added during AI processing, and/or information contributed by the site administrators. Please send Feedback with comments and questions regarding the information presented.

Image
Displays the raw image received. If the image has been processed and the system has detected one or more footprints in the image, a yellow bounding box is drawn around each footprint.
Species, Individual, Sex, Foot
Labels entered by the image contributor either in Epicollect, or edited by WildTrack
Rating, WildTrack Comments
WildTrack rates the quality of the image from a "fit for analysis" perspective on a 5 point scale:
  1. Partial footprint of any quality. A footprint that is incomplete.
  2. Poor Quality. Toe/heel outlines not visible. Hard to see if it's a footprint or something else.
  3. Medium Quality. Toe/heel outlines partially visible.
  4. Good Quality. Toe/heel outlines mostly visible.
  5. Excellent Quality. FIT-usable. Toe/heel outlines clear and undistorted.
A rating of 0 indicates that there was no image received. Any relevant qualitative feedback will be displayed in the WildTrack Comments column. Note that while it is always desirable to have the highest quality image, we are looking to capture as many images as possible (even with lower quality) and as outlined in the success definition section, evaluate how the AI models do with lower quality images.
AI Predictions
Displays the species and individual predicted by the system along with % confidence level. If the confidence level falls below a predefined threshold, Unknown is displayed.

Note that the current model is trained on 13 species (Amur Tiger, Bengal Tiger, Leopard, Cheetah, Puma, Jaguar, African lion, African elephant, Black Rhino, White Rhino, Lowland Tapir, Bongo and Otter). The model tries to assign any image to one of these species.

Recorder Information
The contributor information (along with comments) received with the original image submission.