README.md

Dog_Breed_Classifier

Abstract:

The neural network is the most advanced method at the moment to train the system to do a task. Among them CNN (convolution neural network ) and its derivatives are the most used for training the system to identify and classify images and in computer vision.

Network Used:

Convolution Neural Network (CNN)

Layers in the network :

• cnn.add(Conv2D(filters=128,kernel_size=5,activation='relu',input_shape=[64,64,3]))

• cnn.add(MaxPool2D(pool_size=3,strides=1))

• cnn.add(Dropout(0.2))

• cnn.add(BatchNormalization())

• cnn.add(Conv2D(filters=64,kernel_size=3,activation='relu'))

• cnn.add(MaxPool2D(pool_size=3,strides=1))

• cnn.add(Dropout(0.2))

• cnn.add(BatchNormalization())

• cnn.add(Conv2D(filters=32,kernel_size=3,activation='relu'))

• cnn.add(MaxPool2D(pool_size=3,strides=1))

• cnn.add(Dropout(0.2))

• cnn.add(BatchNormalization())

• cnn.add(Conv2D(filters=32,kernel_size=3,activation='relu'))

• cnn.add(MaxPool2D(pool_size=3,strides=1))

• cnn.add(Dropout(0.2))

• cnn.add(BatchNormalization())

• cnn.add(Conv2D(filters=32,kernel_size=3,activation='relu'))

• cnn.add(MaxPool2D(pool_size=3,strides=1))

• cnn.add(Dropout(0.2))

• cnn.add(BatchNormalization())

• cnn.add(GlobalAveragePooling2D(data_format='channels_last'))

• cnn.add(Dense(units=10,activation='softmax'))

Accurracy Score:

32%. Due to lack of computation power number of epochs had to be set to less, so the accuracy is low.

Dataset : Data set is taken from kaagle

Contributors :

Sanjay urs K P , Jyothi B R, Karthik B S, Nayana , Mohan Kumar K

Licensed To : MIT

Images :