

This directory contains the following two scripts:

   single_instance_object_detection.py*

      This script assumes that an image contains a single meaningful object and 
      its goal is to classify the image based on that object and return its bounding 
      box.  The image is allowed to contain structured artifacts and noise in 
      addition to a single meaningful object.

   multi_instance_object_detection.py

      This script is based on the YOLO logic for recognizing and localizing multiple
      instances of objects in an image.  It also returns the bounding boxes for each 
      of the recognized objects.



The first of these requires the following dataset:

    PurdueDrEvalDataset

and the second requires the following dataset:

    PurdueDrEvalMultiDataset


The main webpage for the RegionProposalGenerator module contains information on how
to download and install the datasets.


For both these scripts, see the Week 8 lecture of my Deep Learning class at Purdue

