Explanation

Image classification Ironman vs Ultron

Dataset:
Ironman and ultron images downloaded by craweler from google.Image_size 224x224 pixels and trainning had each class 40 images and test had each class 10 images.
overviwe:
Create a convolution neural network for classify.
Tools:
Tesnsorflow & keras cnn architecture, Numpy,Jupyternotebook,Data augmentation for better learning.
Model summary:
Convolution neural network(CNN) layers,Maxpooling,flatten,Dense layer
Optimizer : adam (Adaptive moment estimation),loss:binary_crossentropy

Result

Model trainning accuracy: 0.8 to 0.9 (epoch 6-9)learn good,
validation accuracy:
Reached 1.0 by after epoch 5.
Model save in h5 (Hirarchical data format5)

streamlit web app for Model prediction:

User Interface for file upload:

Iron Man:

Ulron: