Dog-Breed-Classifier: Convolutional Neural Nets and Transfer Learning for image classification
Project Overview
In this post, I will be discussing the process of building a pipeline to process real-world, user-supplied images. Given an image of a dog, the algorithm will identify an estimate of the canine’s breed. If supplied with an image of a human, the code will identify the resembling dog breed.
Convolutional Neural Network (CNN) is a special grade of Neural Network that can view and generalize patterns from raw images. It is able to then make use of the different patterns to generalize even more complex patterns. Here the CNN will be used to classify an image into one of the particular breeds.
Making a classifier for Dogs breeds, let alone Humans, are very difficult. Dog breeds have very high intercorrelation between the breeds which makes it difficult to tell breeds apart. That is why this is a good problem to learn more about CNNs.
Problem Statement
We are given various sample images of different dimensions (500 x 500, 1024 x 600, etc.) with no guarantee for the dimensions of the image beforehand. The network is responsible for determining the…
