Pranav 6f3e629cc8 First Commit il y a 6 ans
..
haarcascades 6f3e629cc8 First Commit il y a 6 ans
README.md 6f3e629cc8 First Commit il y a 6 ans
dog_app.html 6f3e629cc8 First Commit il y a 6 ans
dog_app.ipynb 6f3e629cc8 First Commit il y a 6 ans
extract_bottleneck_features.py 6f3e629cc8 First Commit il y a 6 ans
output.png 6f3e629cc8 First Commit il y a 6 ans

README.md

P1: Dog Breed Classifier

In this project, a dog breed recognition algorithm is built using Transfer Learning and Convolutional Neural Networks.

About

A pipeline is built to process real-world, user-supplied images. Given an image of a dog, the developed algorithm provides an estimate of the canine's breed. If supplied with an image of a human, the code identifies the resembling dog breed.

Output

Files

The original repository of this project is https://github.com/udacity/dog-project.

  • dog_app.ipynb – Project notebook.

  • dog_app.html – HTML Export of the project notebook.

  • haarcascades/haarcascade_frontalface_alt.xml – An OpenCV implementation of Haar feature-based cascade classifiers for detecting human faces (source).

  • extract_bottleneck_features.py – Functions to extract the bottleneck features.

Data
  • The training data for this project is available here. This dataset contains 133 different breeds of dogs and is already split into train, test, and validation sets.

  • Human images can be downloaded from this link.

License

Modified MIT License © Pranav Suri

Please note that for this project, @cgearhart, @luisguiserrano are the default owners.