Classification of dogs' breed using Convolutional Neural Networks
Having dataset of human faces and dataset for dog images, the Network is trained to detect the human and dogs in images and gives percentage of the accuracy of the trained model in terms of how many images in the human dataset are detected to be human face and how many images in dog dataset are detected to have dogs. CNN architecture of trained model attains at least 10% accuracy on the test set
Using CNN and the trained model, another job is done which is classifying the dog's breed in each image and if a human is detected, the resembling breed shall be classified and this method is called transfer learning. CNN to Classify Dog Breeds Using Transfer Learning has accuracy more than 70%.
there are some libraries needed for this project including: Python 3.7, Numpy, Pandas, Matplotlib, Torchvision, PyTorch
Download the dog dataset Download the human_dataset
using VGG16 model