search

uptake / autofocus

Deep learning computer vision for classifying wildlife in camera trap images
BSD 3-Clause "New" or "Revised" License
54 stars 30 forks source link

Incorporate a background subtraction step #36

Open gsganden opened 6 years ago

gsganden commented 6 years ago

E.g. cluster empty images, and for each new image identify the closest cluster and apply standard background detection techniques. Could use as a preprocessing step for a classifier or to highlight areas for manual inspection in a GUI.

gsganden commented 6 years ago

One challenge is that the cameras sometimes get jostled and the background sometimes changes, e.g. because a tree falls down. Clustering may or may not address these issues. Image stitching techniques could also be useful.