Demo Team / Wildlife Camera: On Device - North America Public

Wildlife Camera: On Device - North America

Images

About this project

This project implements a simple image classifier for remote wildlife camera animal detection

The data is sourced from the CCT20 benchmark dataset: https://lila.science/datasets/caltech-camera-traps

Additional data from the COCOS image dataset, adding a 'Person' class, has been additionally added to further reduce false positives.

empty.5a130b37-23d2-11e8-a6a3-ec086b02610b
rabbit.59baca0b-23d2-11e8-a6a3-ec086b02610b
coyote.5946b8e2-23d2-11e8-a6a3-ec086b02610b
opossum.59aadd18-23d2-11e8-a6a3-ec086b02610b
opossum.58c34e6e-23d2-11e8-a6a3-ec086b02610b
bird.59c99d12-23d2-11e8-a6a3-ec086b02610b
rabbit.59738963-23d2-11e8-a6a3-ec086b02610b
person.COCO_train2014_000000272228.28netr08.ingestion-577f6767bd-pkb77

Run this model

Scan QR code or launch in browser

Dataset summary

Data collected
15,903 items
Labels
animal, car, empty, person

Project info

Project ID 105811
Project version 3
License BSD 3-Clause Clear
No. of views 114,936
No. of clones 9