This directory contains several uses cases which provide a short narrative describing typical uses of BBoxEE.

Taylor is starting a new camera trapping initiative monitoring Fisher Cats and has just returned from the field with 10 micro SD cards full of images. Taylor is interested in training an object detection model to help automate future sorting and labeling of images from the camera traps.

Taylor has been contacted by a local forest preserve which would like assistance training an object detection model to help automate future sorting and labeling of images from camera traps. The preserve has had a small camera trapping initiative running for the last 10 years and has an archive of 150,000 labeled images. These images are stored in Colorado Parks and Wildlife's CPW Photo Warehouse , a Microsoft Access database.

The objective of this use case / tutoral is to demonstrate a BBoxEE workflow to leverage the power of an existing model to help jump start your annotation of a custom dataset and then fine-tune that model to your project's needs. Fine-tuning allows you to take advantage of an existing model and tailor it to your needs in less time and with less data compared to building a model from scratch. Fine-tuning takes advantage of a technique called transfer learning.