Grace Carter, Alex Michael, Matt Bussing, Rachel McManus
http://vision.stanford.edu/aditya86/ImageNetDogs/.
The original images from the Stanford Dogs Dataset were taken and processed by Alex and Grace. They located the dog in each image and returned a cropped version of the image, that is focused only on the dog.
This data is for my team's final project. We noticed that there were issues with the data. For example, inside toy terriers the photo, n02087046_7214, is an image of something other than a toy terrier. Thus, we are going to fix the inconsistencies and remove photos that will throw off the Machine Learning. For example, a dog wearing a bee costume.
Inconsistencies fixed (removed):
- incorrect breed
- dog in a costume
- multiple dogs
- datacleaning.py: Created by Alex and Grace. Does the cropping.