Animal detection module

Object detection package

Initial setup

• Install Tensorflow object detection API. Alias models/research/object_detection as obj_det. Ensure path is set correctly.
• Set path in file model.py and mavi_object_detection.py. (Search for #TODO)
• Download the pre trained model into the obj_det directory
• Copy misc/mavi_animal_label_map.pbtxt into obj_det/data directory
• The conda environment (saved at misc\conda_environment.yaml) can be loaded as conda env create -f conda_environment.yaml.

To run

Run mavi_object_detection.py from one level up. It runs inference for file at obj_det/test_images/image1.jpg

Model and dataset

Model is fine tuned on Animal dataset (dog and cow) obtained from website.